2010-09-19T00:00:46 *** pc has quit IRC (Remote host closed the connection) 2010-09-19T00:00:46 *** xvinyl has joined #aichallenge 2010-09-19T00:02:23 *** Eiii|laptop has joined #aichallenge 2010-09-19T00:03:22 *** Eiii has quit IRC (Ping timeout: 240 seconds) 2010-09-19T00:03:27 *** McLeopold has quit IRC (Quit: Leaving) 2010-09-19T00:04:24 wait till both bots are good or some time limit, then kill the bots that went over time or didn't respond 2010-09-19T00:08:18 *** Sean_McEligot has quit IRC (Remote host closed the connection) 2010-09-19T00:09:33 *** DirtyKeyboard has quit IRC (Read error: Connection reset by peer) 2010-09-19T00:10:18 *** DirtyKeyboard has joined #aichallenge 2010-09-19T00:10:35 Mathnerd314: not in here 2010-09-19T00:10:55 oh, ok... 2010-09-19T00:11:41 mailing list or forums or issues 2010-09-19T00:15:16 amstan: has that been changed to just a loss of match yet? 2010-09-19T00:15:56 antimatroid1: no, why would it? 2010-09-19T00:16:11 *** DirtyKeyboard has quit IRC (Quit: DirtyKeyboard) 2010-09-19T00:16:18 because it's an extremely strict rule and doesn't really cause any issues to you guys 2010-09-19T00:16:32 it does.. 2010-09-19T00:16:38 because it's kind of pointless that a timeout results in DQ whereas most technical errors are just losses 2010-09-19T00:16:51 such as? 2010-09-19T00:16:57 if someone submits an infinite loop bot he'll just eat resources 2010-09-19T00:17:07 no he wont, he'll die first move every game 2010-09-19T00:17:08 that's why you have the timeout code, sure 2010-09-19T00:17:08 that's quick 2010-09-19T00:17:17 no reason for timeout to be DQ though 2010-09-19T00:17:37 could have max 5 timeouts or something 2010-09-19T00:17:41 discuss this on the mailing list please, i'm just not sure about their reasoning 2010-09-19T00:18:52 is there an open mailing list? the only one I saw (google groups) seemed to be closed to subscriptions 2010-09-19T00:19:29 yeah that's a dev mailing list. seems kind of silly to be closed though 2010-09-19T00:19:33 IMO 2010-09-19T00:19:35 i guess not, issues list then 2010-09-19T00:19:51 jmcarthur: stuff about contestants is also discussed there 2010-09-19T00:19:55 i think that's one of the reasons 2010-09-19T00:19:56 ah 2010-09-19T00:19:57 idk 2010-09-19T00:20:00 makes sense 2010-09-19T00:20:07 although some of the subscribers *are* contestants 2010-09-19T00:20:24 which doesn't seem fair, if that's the reason it's closed in the first place 2010-09-19T00:20:26 that's one of the things that disagree with my reason 2010-09-19T00:20:38 i have no idea, lol 2010-09-19T00:20:40 heh 2010-09-19T00:20:46 :) ok just wanted to make sure I hadn't missed one 2010-09-19T00:20:50 go bother jeff 2010-09-19T00:20:59 Janzert: you can also try the issue list 2010-09-19T00:22:07 yeah, I'm actually waiting on jeff to get time to add me as a dev and setting up a test environment in the mean time 2010-09-19T00:23:14 Janzert: email him, he'll probably get you in 2010-09-19T00:24:45 yeah, that's the communication I've had with him and he said he would be just that he was busy, remind him later :) 2010-09-19T00:30:00 contestbot: later tell j3camero Janzert wants on the mailing list, just a reminder 2010-09-19T00:30:00 amstan: As you wish. 2010-09-19T00:47:01 *** pjreddie_ has joined #aichallenge 2010-09-19T00:50:25 *** pc has joined #aichallenge 2010-09-19T00:50:34 *** xvinyl has quit IRC (Remote host closed the connection) 2010-09-19T00:54:13 omg what!? http://www.benzedrine.cx/planetwars/canvas?game_id=38264 2010-09-19T00:56:05 Top 10 players: McLeopold(3529), kcm1700(3465), EBraun(3430), adumlah(3361), wagstaff(3339), montanalow(3262), Longing(3230), felixcoto(3218), Petrrr(3214), Tracer(3213) 2010-09-19T00:57:02 What happens when a turn passes and a fleet reaches a planet with the equal number of ships on that planet? 2010-09-19T00:57:18 planet population goes to 0, owner is retained 2010-09-19T00:57:27 ty 2010-09-19T00:57:50 *** Theobon has joined #aichallenge 2010-09-19T00:57:51 assuming they have equal numbers *after* growth, that is 2010-09-19T00:58:22 a1k0n: you need to expand more :P 2010-09-19T00:58:38 moar* 2010-09-19T00:58:39 antimatroid1: nah i had that in hand, it just catastrophically ignored the takeover of the 44 at the bottom 2010-09-19T00:58:53 which means theres some kind of giant bug 2010-09-19T00:59:03 So the pop calculation orer goes numfleets += growth rate; add up incoming fleets; repeat 2010-09-19T00:59:14 most of this bot's games just smoothly draw the enemy down once it has a numerical advantage 2010-09-19T00:59:28 Not growth rate after incoming fleets are tallied? 2010-09-19T00:59:31 correct. 2010-09-19T00:59:41 a1k0n, didn't you win last challenge. Iz remember you 2010-09-19T00:59:56 so if a planet has n ships at time t and grows at rate g, and a fleet arrives at g+1, then the new population is n+g - (fleet size) 2010-09-19T01:00:07 er, t+1, not g+1 2010-09-19T01:00:09 *** Queue29 has joined #aichallenge 2010-09-19T01:00:10 blbrown_win3: yup 2010-09-19T01:01:00 he's the kind of 'superstar' around here, if you will 2010-09-19T01:01:38 * a1k0n performance anxiety 2010-09-19T01:02:27 ha, looks like competition will be greater this time 2010-09-19T01:02:48 *** Eiii|laptop has quit IRC (Ping timeout: 272 seconds) 2010-09-19T01:02:52 just gonna slip this in... http://ai-contest.com/visualizer.php?game_id=4534556 2010-09-19T01:03:28 ha yeah. my bot on the official site is terrible 2010-09-19T01:03:50 mine should be ranked higher than like 50th :( 2010-09-19T01:03:54 and my subsequent efforts so far are worse :( 2010-09-19T01:04:30 it's easy to do that 2010-09-19T01:04:44 i'm working on fixing up my backend at the moment then fixing a few thigns i don't want it to do 2010-09-19T01:12:17 a1kon: that's my problem too! I keep making changes and having my old bot destroy the new one. 2010-09-19T01:12:46 my new bot utterly defeats my old one, but is weak sauce to everyone else 2010-09-19T01:14:00 a1k0n: I just stopped testing my bot against itself. It seemed to have no correlation with how it does against the field on benzedrine. 2010-09-19T01:14:29 *** pjreddie_ has quit IRC (Quit: Page closed) 2010-09-19T01:15:56 *** xvinyl has joined #aichallenge 2010-09-19T01:17:19 *** pc has quit IRC (Remote host closed the connection) 2010-09-19T01:17:58 which means there is some serious rock-paper-scissorsing going on 2010-09-19T01:18:12 and we should all have like three independent score components 2010-09-19T01:19:43 http://www.youtube.com/watch?v=iapcKVn7DdY 2010-09-19T01:20:18 i've decided that rock-paper-scissors just means our bots still suck 2010-09-19T01:21:01 jmcarthur: i agree, that behavoir should be taken care of in advanced bots 2010-09-19T01:24:20 disagree. that reasoning just goes "nth level". 2010-09-19T01:26:42 *** McLeopold has joined #aichallenge 2010-09-19T01:29:16 all that mucking around with rocks, papers, and scissors... I find it's just easier to flip a coin 2010-09-19T01:29:29 ...inside a schrodinger box 2010-09-19T01:30:32 ...and the box and I are also inside anoter larger box 2010-09-19T01:30:42 so I don't even know if I flipped the coin or not 2010-09-19T01:31:09 *** clever_name has joined #aichallenge 2010-09-19T01:36:07 awesome... now my bot's being matched with players ranked ~1300 -_- 2010-09-19T01:36:26 the first game of a new submission should really be with someone around the same rank as the old submission 2010-09-19T01:38:09 not very good if you are trying to encourage beginners, either... they get utterly demolished by new submissions working their way up 2010-09-19T01:39:06 jmcarthur and I talked about that before. I think he agreed that the confidence should be increased to max and the average should stay the same. 2010-09-19T01:39:15 yup 2010-09-19T01:39:23 Or he convinced me... 2010-09-19T01:39:47 you mean confidence should be set to minimum? 2010-09-19T01:39:48 I guess the thing to do is play on the tcp server for quick results. 2010-09-19T01:40:19 Well, i supose you could say it that way, but it means your confidence range covers the whole field. 2010-09-19T01:40:29 Meaning a win or loss moves you around quicker. 2010-09-19T01:40:45 ok, yeah... I see what you mean... the confidence interval is increased 2010-09-19T01:40:51 yes 2010-09-19T01:41:42 the current number 5 only played 11 games to get there 2010-09-19T01:41:54 that should only be 5-6 hours 2010-09-19T01:44:44 *** farbe_ has quit IRC (Quit: farbe_) 2010-09-19T01:45:24 unfortunately bayeselo doesn't make this possible 2010-09-19T01:45:52 *** dhartmei has joined #aichallenge 2010-09-19T01:47:41 i think danielvf tested trueskill and said it came out very badly, but i suspect that was not with the trueskill matching algorithm, which i'm pretty sure is the key to getting people to move to their correct scores quickly 2010-09-19T01:48:29 either way, bayeselo is at least very good. that it's so hard to better it is a good sign, at least 2010-09-19T01:53:11 *** HakanD has quit IRC (Ping timeout: 245 seconds) 2010-09-19T01:56:05 Top 10 players: McLeopold(3528), kcm1700(3467), EBraun(3421), wagstaff(3357), adumlah(3347), montanalow(3256), Longing(3248), Petrrr(3239), petesimard(3213), Tracer(3209) 2010-09-19T01:59:17 *** McLeopold has quit IRC (Quit: Leaving) 2010-09-19T02:10:16 *** pacotabaco has quit IRC (Quit: Page closed) 2010-09-19T02:17:17 this contest should come with a warning label 2010-09-19T02:17:30 rwa: agreed 2010-09-19T02:17:41 will ruin lives 2010-09-19T02:22:21 I'm kind of hoping somebody comes along with some kind of uberbot that just destroys me every time. That way I can safely give up. 2010-09-19T02:22:49 but would you give up even then? 2010-09-19T02:23:21 *** farbe_ has joined #aichallenge 2010-09-19T02:23:33 we'd probably come up with something that destroys uberbot 99/100... but then can't touch RandomBot 2010-09-19T02:26:39 there's almost a gambling addiction here... tweak code, pull the submit lever... 2010-09-19T02:26:39 *** dhartmei_ has joined #aichallenge 2010-09-19T02:27:34 come on three cherries against Leopold.5! Daddy needs a new pair of shoes! 2010-09-19T02:27:54 *** hellman has joined #aichallenge 2010-09-19T02:29:09 hahaha 2010-09-19T02:31:45 cherry... cherry... invalid instruction?! nooooooooo I waged my job/health/family life on that one! 2010-09-19T02:34:23 http://challenge.gov/ looks kind of cool 2010-09-19T02:34:50 most of them require bigtime money to compete though 2010-09-19T02:35:03 not just long hours behind a keyboard, which is about all I can muster 2010-09-19T02:38:11 *** jmreardon has quit IRC (Quit: jmreardon) 2010-09-19T02:38:24 alright! I'm on tcp and not timing out! 2010-09-19T02:38:57 sweet, what's your bot name 2010-09-19T02:39:06 sigh-1 and sigh-1A 2010-09-19T02:40:34 they are the almost the same, but different strategies for distributing ships between planets 2010-09-19T02:40:56 nice to start undefeated... 2010-09-19T02:41:46 :) 2010-09-19T02:42:34 I was losing one game, then my opponent made an invalid move 2010-09-19T02:46:43 alright, sigh-1 is at 4th place... I'm coming for you rwa :D 2010-09-19T02:47:29 *** clever_name has quit IRC (Ping timeout: 252 seconds) 2010-09-19T02:47:37 sigh: nice. have fun with dhartmei. he's tough as nails. 2010-09-19T02:48:27 *** farbe_ has quit IRC (Quit: farbe_) 2010-09-19T02:49:29 yeah, it's be interesting once my bot starts getting matched with you guys 2010-09-19T02:49:33 *it'll 2010-09-19T02:51:04 the low sample #1 is always nice though 2010-09-19T02:51:35 sure are a lot of contestants already 2010-09-19T02:51:36 your bot looks quite tough 2010-09-19T02:51:48 seems to have some McLeopold type mannerisms 2010-09-19T02:52:22 haha, awesome... I'll savor my ill-eared #1 2010-09-19T02:52:30 *** boegel has joined #aichallenge 2010-09-19T02:52:32 *earned 2010-09-19T02:52:46 rwa: to me my bot seems more similar to bartwe 2010-09-19T02:52:58 it has characteristics of both 2010-09-19T02:53:09 like this game: http://www.benzedrine.cx/planetwars/canvas?game_id=39288 2010-09-19T02:53:14 you're a hopper like bartwe for sure 2010-09-19T02:53:21 hopper? 2010-09-19T02:54:41 rwa: sigh-1 defeated dhartmei.002: http://www.benzedrine.cx/planetwars/canvas?game_id=39408 :D 2010-09-19T02:55:14 you're the real deal 2010-09-19T02:55:32 :) 2010-09-19T02:55:37 nice. did you submit that on the contest site? 2010-09-19T02:55:44 you hop from planet to planet rather than go straight from A->B 2010-09-19T02:55:48 *** pc has joined #aichallenge 2010-09-19T02:55:49 *** xvinyl has quit IRC (Remote host closed the connection) 2010-09-19T02:55:55 currently sigh-1A is on the contest site 2010-09-19T02:56:06 Top 10 players: McLeopold(3519), EBraun(3408), adumlah(3394), kcm1700(3358), wagstaff(3355), George_Whiteside(3355), Macuyiko(3253), montanalow(3236), Longing(3214), Tracer(3209) 2010-09-19T02:56:11 raw: ah, that's sigh-1A's redistribution strategy 2010-09-19T02:56:48 whoa, top 10? when did that happen 2010-09-19T02:56:59 wh1teside: nice :) 2010-09-19T02:57:55 *** hellman has quit IRC (Remote host closed the connection) 2010-09-19T02:58:40 whoa, you beat McLeopold.5, too 2010-09-19T02:58:45 handily 2010-09-19T03:00:44 oh wow... just watched that game 2010-09-19T03:01:24 and bartwe 2010-09-19T03:02:03 *** JCS^ has quit IRC (Quit: \x00) 2010-09-19T03:02:19 yay... elo over 700 2010-09-19T03:05:51 *** Naith has joined #aichallenge 2010-09-19T03:06:54 *** farbe_ has joined #aichallenge 2010-09-19T03:07:20 ah, you're not unbeatable 2010-09-19T03:07:27 :P 2010-09-19T03:07:41 that game was a good one for me, exposed a big weakness 2010-09-19T03:11:06 *** boegel has quit IRC (Ping timeout: 245 seconds) 2010-09-19T03:14:10 *** boegel has joined #aichallenge 2010-09-19T03:15:32 *** farbe_ has quit IRC (Read error: Connection reset by peer) 2010-09-19T03:16:02 *** boegel has quit IRC (Client Quit) 2010-09-19T03:17:24 *** boegel has joined #aichallenge 2010-09-19T03:24:43 *** syntaxglitch has quit IRC (Ping timeout: 265 seconds) 2010-09-19T03:24:53 alright, sigh, you've got me running scared. I guess I'm gonna have to think about my strategy rather than just tweak parameters and watch leaderboards now. :P 2010-09-19T03:25:25 but you still haven't played me yet :P 2010-09-19T03:27:07 *** blbrown_win3_ has joined #aichallenge 2010-09-19T03:28:02 sigh, I know, I've got four bots trolling for you with no luck 2010-09-19T03:28:03 frankly it doesn't seem like a bad strategy at the moment :) 2010-09-19T03:28:52 it would be awesome if the server allowed you to request matches 2010-09-19T03:31:04 *** blbrown_win3 has quit IRC (Ping timeout: 276 seconds) 2010-09-19T03:31:21 add some special moves and fatalities, and we have the makings of the next out-of-nowhere arcade hit 2010-09-19T03:32:05 :) 2010-09-19T03:33:18 *** boegel has quit IRC (Quit: This computer has gone to sleep) 2010-09-19T03:33:49 picked up sigh-1A 2010-09-19T03:34:18 hoping for a timeout...lol 2010-09-19T03:34:18 awesome 2010-09-19T03:34:21 :D 2010-09-19T03:34:53 I'm ahead now 2010-09-19T03:35:14 I'd be surprised if you didn't win 2010-09-19T03:35:31 yeah, I've got this one 2010-09-19T03:35:46 up 300-70 2010-09-19T03:35:49 it's alright, I've got 51 more - maybe one of those can take you 2010-09-19T03:36:55 good game 2010-09-19T03:37:00 *** sigh is now known as sigh-afk 2010-09-19T03:37:05 see you on top of the leaderboard in the morning... 2010-09-19T03:37:28 rwa: hahaha 2010-09-19T03:38:18 *** hellman has joined #aichallenge 2010-09-19T03:38:27 somebody likes the "knobs" in their AI 2010-09-19T03:40:09 uh oh 2010-09-19T03:40:18 now you're sicking sigh-1 on me 2010-09-19T03:40:26 *** xvinyl has joined #aichallenge 2010-09-19T03:40:32 maybe I'll get a better board this time 2010-09-19T03:40:38 *** pc has quit IRC (Remote host closed the connection) 2010-09-19T03:40:42 I suck at those close-in starting position boards 2010-09-19T03:41:37 *** delt0r___ has joined #aichallenge 2010-09-19T03:41:47 are you running away with it again? I don't watch progress, only win/lose 2010-09-19T03:42:16 sigh 2010-09-19T03:42:22 rwa: my bot sucks on starting close together too 2010-09-19T03:42:31 currently rewriting a lot of things trying to fix such mistakes 2010-09-19T03:43:37 *** delt0r_ has quit IRC (Ping timeout: 276 seconds) 2010-09-19T03:43:40 I had a mod that did well against ragebot on map 30, but it sucked otherwise 2010-09-19T03:44:28 ahhh map30, that's a number I have memorized 2010-09-19T03:45:04 dangit 2010-09-19T03:45:07 sigh beats me again 2010-09-19T03:46:41 close battle though 2010-09-19T03:48:56 *** kcn1700 has joined #aichallenge 2010-09-19T03:49:52 0/3, I guess I call it a night 2010-09-19T03:52:55 woo 2010-09-19T03:52:57 1/4 2010-09-19T03:56:06 Top 10 players: adumlah(3551), McLeopold(3477), EBraun(3407), kcm1700(3348), George_Whiteside(3344), wagstaff(3338), Macuyiko(3286), montanalow(3264), Petrrr(3238), dmj111(3205) 2010-09-19T03:56:13 oh my name is here 2010-09-19T03:56:47 highlighted beep wakes me. 2010-09-19T04:00:24 how about introducing 100 additional maps on the tcp server, generated by the official generator? 2010-09-19T04:01:00 this would give us some preparation ahead of the official server 2010-09-19T04:03:20 as far as I'm concerned, the bigger the map pool the better 2010-09-19T04:03:41 ok 2010-09-19T04:05:31 they are in, map101-200.txt, you can view them through the web interface (getmaps) 2010-09-19T04:05:51 oh nice, thank you very much sir 2010-09-19T04:06:20 i'm off for a couple of hours, i'll be back later, cya 2010-09-19T04:06:48 *** HakanD has joined #aichallenge 2010-09-19T04:07:21 see ya 'round 2010-09-19T04:07:49 *** dhartmei has quit IRC (Quit: This computer has gone to sleep) 2010-09-19T04:09:43 *** Kastra has joined #aichallenge 2010-09-19T04:09:54 *** mog_ has joined #aichallenge 2010-09-19T04:16:24 contestbot: what's up 2010-09-19T04:16:25 wh1teside: An error occured while trying to show the previous error. 2010-09-19T04:19:03 *** HakanD has quit IRC (Quit: Colloquy for iPhone - http://colloquy.mobi) 2010-09-19T04:20:33 contestbot: later tell rwa I got your number -- http://www.benzedrine.cx/planetwars/canvas?game_id=40002 ... except we both called at the same time and got a busy signal. 2010-09-19T04:20:33 wh1teside: Job's done. 2010-09-19T04:26:26 *** TapewormGalactos has joined #aichallenge 2010-09-19T04:26:35 *** Theobon has quit IRC (Quit: Leaving.) 2010-09-19T04:30:38 *** xvinyl has quit IRC (Remote host closed the connection) 2010-09-19T04:30:42 *** cabron has joined #aichallenge 2010-09-19T04:31:03 *** xvinyl has joined #aichallenge 2010-09-19T04:44:43 *** Cyndre_ has quit IRC (Ping timeout: 240 seconds) 2010-09-19T04:46:39 *** Cyndre has joined #aichallenge 2010-09-19T04:51:31 *** sigh-afk is now known as sigh 2010-09-19T04:51:40 *** Cyndre has quit IRC (Ping timeout: 276 seconds) 2010-09-19T04:53:42 *** Cyndre has joined #aichallenge 2010-09-19T04:54:52 *** amstan has quit IRC (Ping timeout: 245 seconds) 2010-09-19T04:55:23 *** xvinyl has quit IRC (Remote host closed the connection) 2010-09-19T04:55:36 *** xvinyl has joined #aichallenge 2010-09-19T04:56:07 Top 10 players: McLeopold(3467), adumlah(3450), Macuyiko(3364), kcm1700(3350), montanalow(3314), wagstaff(3290), dmj111(3262), George_Whiteside(3256), Longing(3250), Tracer(3227) 2010-09-19T04:56:49 hmmm... my enemy planet evaluation needs a lot of work, my bot keeps attacking stupid things 2010-09-19T04:59:39 *** l4u_ has joined #aichallenge 2010-09-19T05:01:46 *** l4u_ has quit IRC (Read error: Connection reset by peer) 2010-09-19T05:03:22 *** l4u has quit IRC (Ping timeout: 276 seconds) 2010-09-19T05:04:52 *** xvinyl has quit IRC (Ping timeout: 245 seconds) 2010-09-19T05:05:06 *** xvinyl has joined #aichallenge 2010-09-19T05:07:28 sigh: mine too 2010-09-19T05:08:46 5 growth rate planet with 7 neutrals right next me --and instead it attack the enemy across the map with 100 ship in it and only growth rate of 2 2010-09-19T05:08:54 I am assuming i have a bug 2010-09-19T05:09:26 That planet must of had valuable resources 2010-09-19T05:09:41 Or your fleets just wanted to go on shore leave 2010-09-19T05:10:08 Insubordinate pilots more like 2010-09-19T05:12:27 wow, everybody got significantly stronger 2010-09-19T05:14:48 bartwe: tell me about it, i think we also had a time when a lot of people were asleep last night 2010-09-19T05:18:09 for me it was fixing my timeout problems, turned out switching from wireless to wired did it :P 2010-09-19T05:18:12 bartwe: the tcp server is hardcore, it changes people :) 2010-09-19T05:18:40 yeah clearly 2010-09-19T05:18:48 hmz, my ag of tricks is running dry 2010-09-19T05:19:28 * sigh steals the rest of what's in bartwe's ag 2010-09-19T05:19:56 b+ 2010-09-19T05:21:15 *** hellman has quit IRC (Remote host closed the connection) 2010-09-19T05:21:52 *** choas has joined #aichallenge 2010-09-19T05:21:57 i'm currently working on a fairly updated version 2010-09-19T05:22:07 hope i get it done tonight, wont have time for botting during the week 2010-09-19T05:22:24 considering you both seem to have me currently 2010-09-19T05:26:11 the main contest is stronger then benzedrine it seems 2010-09-19T05:26:25 or people have multiple usernames on the main contest 2010-09-19T05:27:05 *** Any_Key has quit IRC (Remote host closed the connection) 2010-09-19T05:27:45 *** blkperl_ has joined #aichallenge 2010-09-19T05:27:51 *** millertime has quit IRC (Read error: Operation timed out) 2010-09-19T05:27:54 *** blkperl has quit IRC (Read error: Connection reset by peer) 2010-09-19T05:28:35 *** millertime has joined #aichallenge 2010-09-19T05:29:10 agreed on that 2010-09-19T05:34:35 oh well, tcp is still better than just testing against my own bots 2010-09-19T05:42:33 sigh-1A is doing better against the higher ranks... now up to #6 2010-09-19T05:44:04 bartwe: thinking about it more, it's not necessarily that the main contest bots are stronger, just that there are more of them and there is no dominating strategy yet 2010-09-19T05:45:11 i still see 40 bots aove me 2010-09-19T05:45:14 *** xvinyl has quit IRC (Remote host closed the connection) 2010-09-19T05:45:26 where there first there were 10, against a weaker version 2010-09-19T05:45:36 or i'm overfit against benzedrine 2010-09-19T05:45:41 *** xvinyl has joined #aichallenge 2010-09-19T05:48:07 still, there is a lot of variation... montanalow is 4th on the main site and he has 2 losses to people ranked 21 and 22 2010-09-19T05:49:11 hell, I can't beat ragebot on 100% of the starter maps 2010-09-19T05:50:36 this makes me a ragebot 2010-09-19T05:50:37 http://www.engadget.com/2010/09/18/intel-wants-to-charge-50-to-unlock-stuff-your-cpu-can-already-d/ 2010-09-19T05:53:31 that's quite common in a lot of products (although it's not as obvious)... and it does make sense 2010-09-19T05:54:06 its worse then drm, as the 'no production cost' and 'copies are cheap' arguments does not hold 2010-09-19T05:54:30 simple market segmentation that undermines technological progress 2010-09-19T05:55:12 bartwe: why? 2010-09-19T05:56:04 i mean, you would be perfectly happy if they actually manufactured two separate chips 2010-09-19T05:56:07 Top 10 players: adumlah(3478), McLeopold(3461), kcm1700(3346), montanalow(3336), Macuyiko(3328), wagstaff(3286), George_Whiteside(3277), dmj111(3236), Tracer(3233), paulus(3230) 2010-09-19T05:56:22 you don't think its repulsive for a producer to increase cost (adding (un)locking support) to produce an inferior product ? 2010-09-19T05:59:07 not necessarily inferior to all people... those who only need the reduced power can get it at a cheaper price 2010-09-19T05:59:26 when normally they would not have been able to buy the product at all 2010-09-19T06:01:31 these aren't binned chips, they have fully functional parts that is allready in the price of the product 2010-09-19T06:03:17 *** vikhyat has joined #aichallenge 2010-09-19T06:03:21 what if "the price of the product" took into account market segmentation 2010-09-19T06:03:27 sigh: the point is that with a "competative" market place... they couldn't do that 2010-09-19T06:03:49 *** Yoshi-TS4 has quit IRC (Quit: Minus One IRC) 2010-09-19T06:04:11 But due the intel doing things like "we pay you to *not* stock amd" means there isn't competion 2010-09-19T06:04:21 oh well, I'm against that 2010-09-19T06:04:26 however ARM is comming... i hope... to the desktop 2010-09-19T06:04:59 even with via/amd/intel is still all x86 2010-09-19T06:05:33 but in regard to market segmentation, it's strikes me as funny that if intel actually manufactured two separate chips then people would not complain even though tht might be MORE expensive 2010-09-19T06:05:50 good night everyone 2010-09-19T06:06:25 sigh: But creating *artifical* segmentation is stupid... sure it can work for a monopolistic companay... 2010-09-19T06:06:41 *** phreeza has joined #aichallenge 2010-09-19T06:06:46 but that doesn't make it fair, good or even legal (in some places) 2010-09-19T06:07:15 what do you mean that it only works for a monopolistic company? 2010-09-19T06:07:23 they make more money 2010-09-19T06:07:40 from intels point of view... sure why not 2010-09-19T06:07:52 why can't it work for a company which is not a monopoly? 2010-09-19T06:08:13 because they other company can trivily undercut them 2010-09-19T06:08:51 not if they can't make a profit by doing so 2010-09-19T06:10:31 if the price is below the high end price, you run the risk of not being profitable enough, if the price is above the low end price then you cut out potential buyers 2010-09-19T06:10:54 *** xvinyl has quit IRC (Remote host closed the connection) 2010-09-19T06:11:16 *** xvinyl has joined #aichallenge 2010-09-19T06:16:35 *** RainCT has joined #aichallenge 2010-09-19T06:16:39 sigh: your missing the point. Its *not* below the price point --the cost is the same to intel no matter *what* version you buy --So the low priced versions are not loss leaders.. (note that selling below cost is also anti competiive in most countries) 2010-09-19T06:16:53 any ot... so it doesn't really matter 2010-09-19T06:18:16 anyway, the reason I can't get worked up over is that to an buyer there is no difference from two separately manufactured products... maybe I'm wrong 2010-09-19T06:18:48 *** blkperl_ is now known as blkpelr 2010-09-19T06:18:50 *** blkpelr is now known as blkperl 2010-09-19T06:19:48 sigh: no differnce other than you are paying an inflated cost because there is no competion because of intels underhanded methods 2010-09-19T06:20:11 paying more than you need to *is a difference to this consumer* 2010-09-19T06:22:36 I'm just talking about market segmentation in general. you may be right about intel... I don't know 2010-09-19T06:30:36 *** cabron has quit IRC (Quit: Page closed) 2010-09-19T06:32:18 *** justin_pdx has joined #aichallenge 2010-09-19T06:32:55 sigh: they go a big fine for anti competitive practices both in the us and the eu 2010-09-19T06:32:59 they got 2010-09-19T06:34:22 delt0r___: was that in relation to market segmentation? 2010-09-19T06:34:31 *** espes has quit IRC (Ping timeout: 272 seconds) 2010-09-19T06:34:52 To paying people off to not stock AMD 2010-09-19T06:35:44 but that's got nothing to do with market segmentation... I wasn't arguing that intel weren't anti-competitive 2010-09-19T06:35:51 *** pc has joined #aichallenge 2010-09-19T06:35:55 *** xvinyl has quit IRC (Remote host closed the connection) 2010-09-19T06:36:08 or is there a link I am missing? 2010-09-19T06:36:11 sigh: But it does--because you can't do what intel want to do with competition. 2010-09-19T06:36:53 any its a bit ot... perhaps we should talk borg 2010-09-19T06:37:10 ok, ill stop now :) 2010-09-19T06:37:15 :) 2010-09-19T06:37:44 sigh do you have a bot on the main ladder 2010-09-19T06:38:20 yes, it's called sighbot... it's doing awesomely at #182 2010-09-19T06:40:34 main ladder takes ages to go up the ranks when you upload a new bot :( 2010-09-19T06:42:38 *** justin_pdx has quit IRC (Quit: justin_pdx) 2010-09-19T06:44:32 what about you delt0r___? 2010-09-19T06:44:43 Cool... 2010-09-19T06:44:57 I an just going in and out of the tcp server at this stage. 2010-09-19T06:45:07 what's your name on tcp? 2010-09-19T06:45:18 Once i have a basic hurstic bot i will add it to the main server 2010-09-19T06:45:26 sigh: deltor 2010-09-19T06:45:29 delt0r 2010-09-19T06:45:59 ah, it's not showing any games for you 2010-09-19T06:46:16 I haven't gone on since friday 2010-09-19T06:46:22 refactor 2010-09-19T06:46:25 i see 2010-09-19T06:46:28 you should upload something to the main site, if only to make sure what you are doing compiles and all 2010-09-19T06:46:38 I had some trouble with that on the main site 2010-09-19T06:47:35 yea 2010-09-19T06:47:43 hoping java 1.6 is supported 2010-09-19T06:48:07 well, the game engine doesn't run on anything else so I would hope so :P 2010-09-19T06:48:36 have java 1.5 on my mac and it didn't like that 2010-09-19T06:48:36 the server uses python IIRC 2010-09-19T06:48:37 *** jesusaurus1 has joined #aichallenge 2010-09-19T06:48:40 yup 2010-09-19T06:48:49 macs are java 1.6 arn't they? 2010-09-19T06:48:56 *** jesusaurus has quit IRC (Read error: Connection reset by peer) 2010-09-19T06:49:04 I deploy my work code to macs with no problems 2010-09-19T06:49:38 it's 1.5 on mine, so I just gave up and ran it on my ubuntu desktop 2010-09-19T06:50:00 nothings particularly up to date on my mac 2010-09-19T06:50:07 *** phreeza has quit IRC (Quit: Computer has gone to sleep) 2010-09-19T06:50:12 it's still OSX 10.4 :D 2010-09-19T06:50:20 ic --yea if you update you get java 1.6 2010-09-19T06:50:58 *** boegel has joined #aichallenge 2010-09-19T06:51:39 *** boegel has quit IRC (Client Quit) 2010-09-19T06:53:18 *** espes has joined #aichallenge 2010-09-19T06:56:09 Top 10 players: adumlah(3523), McLeopold(3461), Macuyiko(3379), kcm1700(3368), montanalow(3321), George_Whiteside(3273), wagstaff(3269), dmj111(3238), paulus(3227), Tracer(3216) 2010-09-19T06:59:08 *** atmz443 has quit IRC (Ping timeout: 276 seconds) 2010-09-19T07:00:42 *** smellyhippy has quit IRC (Ping timeout: 240 seconds) 2010-09-19T07:02:15 *** phreeza has joined #aichallenge 2010-09-19T07:05:53 *** smellyhippy has joined #aichallenge 2010-09-19T07:07:51 *** TapewormGalactos has left #aichallenge ("Ex-Chat") 2010-09-19T07:22:25 *** arik181 has quit IRC (Read error: Connection reset by peer) 2010-09-19T07:23:38 *** atmz443 has joined #aichallenge 2010-09-19T07:26:40 *** arik181 has joined #aichallenge 2010-09-19T07:27:54 sigh: http://www.benzedrine.cx/planetwars/canvas?game_id=41706 2010-09-19T07:27:56 grrrr 2010-09-19T07:28:07 *** phreeza has quit IRC (Read error: Connection reset by peer) 2010-09-19T07:29:18 bartwe: haha... again 2010-09-19T07:29:57 8-alt2 should be immunized against that 2010-09-19T07:30:21 cool 2010-09-19T07:30:34 I also need to work on something similar: http://www.benzedrine.cx/planetwars/canvas?game_id=41665 2010-09-19T07:31:55 dhartmei changed tactics 2010-09-19T07:32:18 new bot? 2010-09-19T07:32:26 is everyone using hardcoded strategies or are there some bots with search techniques out there? 2010-09-19T07:32:40 depends on what you call search 2010-09-19T07:33:02 I don't do minimax style search 2010-09-19T07:33:33 I do do small searches for optimal departure day and fleet sizes though 2010-09-19T07:34:05 i'm using search techniques. 2010-09-19T07:34:06 well, optimal by my own crappy heuristics :P 2010-09-19T07:34:06 thats one step i'm missing 2010-09-19T07:34:11 delayed departure 2010-09-19T07:34:19 it's on 4th place :-) 2010-09-19T07:34:35 kcn1700: awesome :) 2010-09-19T07:34:41 kcn1700: what kind of search ? 2010-09-19T07:34:52 but actually 95% of my code are heuristics :-) 2010-09-19T07:34:58 *** Macuyiko has joined #aichallenge 2010-09-19T07:35:59 my code is made of heuristics, greedy searches, and crappy predictions 2010-09-19T07:36:09 takes zero cpu time so far 2010-09-19T07:36:18 bartwe: sounds like mine :) 2010-09-19T07:36:33 sigh-1A plays very similar to your bot 2010-09-19T07:36:44 in some games the play is almost symetrical 2010-09-19T07:36:49 yeah i see that people are starting to steal my minimal spanning tree 2010-09-19T07:37:14 I don't use a minimal spanning tree 2010-09-19T07:37:29 dhartmei has somethign different with the same effect it seems, feeding towards the enemy 2010-09-19T07:37:29 any similarity is an emergent effect :) 2010-09-19T07:37:40 yeah, that's what I do 2010-09-19T07:38:18 *** phreeza has joined #aichallenge 2010-09-19T07:38:21 sigh-1 was my first attempt to implement it, but had a bug which somehow made it do better 2010-09-19T07:38:37 well, that's how things go 2010-09-19T07:39:36 you eat me pretty consistently 2010-09-19T07:39:37 i'm gonna add some flow networks to my code :P 2010-09-19T07:39:42 i like this one :) http://www.benzedrine.cx/planetwars/canvas?game_id=41779 2010-09-19T07:40:28 haha 2010-09-19T07:40:39 a very small numerical superiority finally wins it for me 2010-09-19T07:40:42 here's one from earlier with a lot of symetry: http://www.benzedrine.cx/planetwars/canvas?game_id=39288 2010-09-19T07:41:12 yeah, you have the numbers here 2010-09-19T07:41:21 mine fails to grab the neutral early enough 2010-09-19T07:41:36 *** hellman has joined #aichallenge 2010-09-19T07:41:48 in the game you linked it was dead even until move 33 2010-09-19T07:41:50 bartwe: McLeopold seems to make excellent use of the MST as well 2010-09-19T07:42:45 i don't think my ot has any competition chance 2010-09-19T07:43:03 if i see how much stronger things have gotten this weekend alone 2010-09-19T07:43:10 If you have a // comment it gets copied over with the rest of it, right? 2010-09-19T07:43:26 javagamer: what do you mean? 2010-09-19T07:43:41 *** E9c has joined #aichallenge 2010-09-19T07:43:52 Macuyiko: hmmm... McLeopold doesn't look like he's using MST to me... do you have a game which demonstrates it? 2010-09-19T07:44:25 sigh: just a second, I'll look for one 2010-09-19T07:44:27 bartwe: early days still... you get discouraged too easily :P 2010-09-19T07:47:16 sigh: take this one for example: http://www.benzedrine.cx/planetwars/canvas?game_id=41473 You're right... it's not really a MST, but he does like to send his growthrate of ships to his own planets every turn, after his initial expansion, which takes up about half the game 2010-09-19T07:48:10 whereas bartwe seems to execute one turn of expansion (take over close neutral planets), and then immediatly starts building a MST 2010-09-19T07:49:02 yeah, McLeopold is nothing like an MST 2010-09-19T07:49:12 sigh-1A is more MST-like and doesn't use an MST :P 2010-09-19T07:49:48 you're right, but they all seem to be building 'networks' between their planets of some kind 2010-09-19T07:50:10 yah, that's a natural consequence of having to redistribute your ships 2010-09-19T07:50:41 yeah, I realized that too late and had to add sloppy code to my bot, I'll need to start over again in a week or see 2010-09-19T07:50:43 so* 2010-09-19T07:51:26 Uh, I'm a noob and maybe this is a dumb question, but what's an MST? Quick forum search didn't turn anything up. 2010-09-19T07:51:36 minimum spanning tree 2010-09-19T07:52:12 you can read about it on wikipedia 2010-09-19T07:52:40 *** Olathe has joined #aichallenge 2010-09-19T07:52:57 i think a next step for me would be to make the mst emergent of the algorithm instead of a hardcoded thing 2010-09-19T07:53:10 I am. I just wasn't familiar with the term, thanks for the answer and sorry for the interruption. 2010-09-19T07:53:12 or atleast apply more prediction to maximize flow 2010-09-19T07:53:21 E9c: no worries, questions always welcome :) 2010-09-19T07:53:52 E9c: geh, don't worry about it, i didn't even know it was a MST when i was implementing it 2010-09-19T07:54:05 bartwe: to be honest, I can't see the benefit of MST for this... I rejected it out of hand 2010-09-19T07:54:26 sigh: maybe thats true now 2010-09-19T07:55:03 sigh: earlier it was a very strong element due to 1. concentrating units near enemies 2. maximizing flexibility to redeploy them as defense 2010-09-19T07:55:20 *** vikhyat has quit IRC (Remote host closed the connection) 2010-09-19T07:55:32 yeah... any redistribution at all with those goals is good 2010-09-19T07:55:37 anyways, i don't do defense, i do preemptive recaptures 2010-09-19T07:55:40 bartwe: exactly, I like how some players rebalance the flow directions to support planets which become under attack 2010-09-19T07:55:42 having ships stuck at the back is a waste 2010-09-19T07:56:08 Top 10 players: adumlah(3523), McLeopold(3453), kcm1700(3434), Macuyiko(3374), montanalow(3319), wagstaff(3265), dmj111(3239), paulus(3235), golfin(3211), Tracer(3206) 2010-09-19T07:56:13 once too many of my planets become under attack, I get stuck in trying to defend them instead of going all in 2010-09-19T07:58:14 whuh ? i don't recognize these tactics at all after move 30 orso http://www.benzedrine.cx/planetwars/canvas?game_id=41952 2010-09-19T07:58:18 total fubar 2010-09-19T07:58:24 yay --I can now play as a human 2010-09-19T07:59:05 bartwe: what do you mean? 2010-09-19T07:59:40 complete and utter misprediction of the EV of that +1 neutral next to the enemy 2010-09-19T07:59:56 which escalates into a huge turf war 2010-09-19T08:00:35 ah yes 2010-09-19T08:00:53 are you saying you should not have attacked it at all, or attacked a move later? 2010-09-19T08:01:43 *** DirtyKeyboard has joined #aichallenge 2010-09-19T08:02:02 i think not at all 2010-09-19T08:02:04 oh sorry, not attacked it at all 2010-09-19T08:02:19 I didn't realise for a sec you were attacking from so far away 2010-09-19T08:02:38 red was silly, he wasted 18 units 2010-09-19T08:02:44 i stubbed out the oppertunity cost of long attacks because i couldn't find the right scale and unit 2010-09-19T08:02:57 stubbed out? 2010-09-19T08:03:06 return 0; 2010-09-19T08:03:11 ah :) 2010-09-19T08:03:26 not having control over units should be a cost 2010-09-19T08:03:55 I'd be more concerned that you launched that fleet at all on move 16 2010-09-19T08:03:57 *** meow__ has quit IRC (Ping timeout: 252 seconds) 2010-09-19T08:04:09 *15 2010-09-19T08:04:09 or maybe you see it as expending units, but then there should be a cost in giving the opponent information with such a long window 2010-09-19T08:04:25 yes, most definetly 2010-09-19T08:04:47 that's why I try to delay my launches as much as possible, especially with coordinated attacks 2010-09-19T08:05:02 yeah at move 16 it allready looks like a stupid move, the recapture would be near instant allready 2010-09-19T08:05:06 if you watch carefully, my coordinated attacks mostly reach the destination at the same time 2010-09-19T08:05:09 *** Apophis_ has joined #aichallenge 2010-09-19T08:05:31 i don't think this move can even have an ev+ 2010-09-19T08:05:34 must be a bug 2010-09-19T08:05:44 yeah, I would say so 2010-09-19T08:06:02 the +2 a few moves lower is a recapture guarantee 2010-09-19T08:06:03 atm, my code explicitly dis-allows attacking neutrals closer to the enemy 2010-09-19T08:06:32 ouch, you capture the 32 as well 2010-09-19T08:06:44 really ? 2010-09-19T08:06:54 hmz, inverted condition somewhere i'm guessing :) 2010-09-19T08:06:55 move 21 2010-09-19T08:06:58 *41 2010-09-19T08:07:25 has to be a bug, I can't see how any sane algorithm would generate those moves :P 2010-09-19T08:07:35 moah, 41 may make sense 2010-09-19T08:08:12 how??? 2010-09-19T08:08:48 red spent like 4 units to take it back 2010-09-19T08:09:32 *** nann has quit IRC (Ping timeout: 272 seconds) 2010-09-19T08:09:40 you lose 29 ships, and give red a planet 2010-09-19T08:10:55 *** E9c has quit IRC (Quit: Page closed) 2010-09-19T08:13:40 If you launch an attack on an enemy planet 1 turn away does growth apply before your ships arrive? 2010-09-19T08:14:49 from what I've been told that are no planets 1 turn away 2010-09-19T08:15:13 but... the order of actions is detailed here: http://www.ai-contest.com/specification.php#game-state-update 2010-09-19T08:15:18 Thanks 2010-09-19T08:15:27 *** xvinyl has joined #aichallenge 2010-09-19T08:15:52 *** ntroPi has joined #aichallenge 2010-09-19T08:16:38 *** pc has quit IRC (Remote host closed the connection) 2010-09-19T08:18:24 *** Yoshi-TS4 has joined #aichallenge 2010-09-19T08:24:17 *** atmz443 has quit IRC (Read error: Connection reset by peer) 2010-09-19T08:24:28 does this map come out half blank for everyone else: http://www.benzedrine.cx/planetwars/canvas?game_id=42115 2010-09-19T08:25:31 *** atmz443 has joined #aichallenge 2010-09-19T08:25:45 *** misato has joined #aichallenge 2010-09-19T08:25:49 hi there 2010-09-19T08:26:13 hi 2010-09-19T08:26:30 i'm trying to register in the forums but it allways says "incorrect captcha" 2010-09-19T08:26:32 :( 2010-09-19T08:26:42 I've got a newbie question 2010-09-19T08:27:14 Can't really help you with the capture, but ask any question you like 2010-09-19T08:27:24 *captcha 2010-09-19T08:27:50 *** kcn1700 has quit IRC (Quit: ¡¬(^o^)£¯) 2010-09-19T08:28:06 i'ts a very newbie question 2010-09-19T08:28:09 :P 2010-09-19T08:28:22 i've got the "timeout" problem with the java starter package 2010-09-19T08:28:44 i've read this threadhttp://ai-contest.com/forum/viewtopic.php?f=18&t=534 2010-09-19T08:28:50 and tried everything in there 2010-09-19T08:28:55 but with not solution :( 2010-09-19T08:29:19 it times out on the first move? 2010-09-19T08:29:32 yep 2010-09-19T08:29:48 it says "player 1 timeout" "player2 timeout" 2010-09-19T08:29:54 i'm over Mac OS 10.6 2010-09-19T08:30:18 what version of java? 2010-09-19T08:30:22 java version "1.6.0_20" Java(TM) SE Runtime Environment (build 1.6.0_20-b02-279-10M3065) Java HotSpot(TM) 64-Bit Server VM (build 16.3-b01-279, mixed mode) 2010-09-19T08:30:28 i was looking for it :P 2010-09-19T08:30:53 hmmm... java version looks ok 2010-09-19T08:31:08 usually timeouts like that mean that the bot is throwing an error 2010-09-19T08:31:17 have you tried manual input to bots? 2010-09-19T08:31:31 no 2010-09-19T08:31:41 i've just downloaded the starter package 2010-09-19T08:31:55 and tried with the example bots, just to see if everything was ok 2010-09-19T08:32:09 the example bots don't work? 2010-09-19T08:32:17 exactly 2010-09-19T08:32:25 what's the command you are running? 2010-09-19T08:32:25 don't work any of them, it sais timeout 2010-09-19T08:32:27 did you use the new .jar files from the thread you cited? 2010-09-19T08:32:42 yes i've used them 2010-09-19T08:33:03 hmm that worked for me ... had the same problem in the beginnen :/ 2010-09-19T08:33:15 well, let's try again 2010-09-19T08:33:20 just in case i missed something 2010-09-19T08:33:20 xD 2010-09-19T08:33:24 :-D 2010-09-19T08:33:44 *** boegel has joined #aichallenge 2010-09-19T08:34:05 *** rtaycher has joined #aichallenge 2010-09-19T08:36:13 nah 2010-09-19T08:36:18 it crashes the same way 2010-09-19T08:36:19 WARNING: player 1 timed out. WARNING: player 2 timed out. Turn 1 Draw! 2010-09-19T08:36:27 misato: what's the exact command you are running? 2010-09-19T08:36:31 anyone else have find the c# package broken? or is it something on my end?(note I downloaded package a while ago) 2010-09-19T08:36:32 java -jar tools/PlayGame-1.2.jar maps/map7.txt 1000 1000 log.txt "java -jar example_bots/RandomBot.jar" "java -jar example_bots/DualBot.jar" | java -jar tools/ShowGame-1.2.jar 2010-09-19T08:36:35 this one 2010-09-19T08:37:00 that looks right :S 2010-09-19T08:37:09 what happens if you run: java -jar example_bots/RandomBot.jar 2010-09-19T08:37:12 by itself 2010-09-19T08:37:14 hum 2010-09-19T08:37:16 let's see 2010-09-19T08:37:44 paste him contents of a map file and "go" 2010-09-19T08:38:01 what? 2010-09-19T08:38:23 sigh: it does (apparently) nothing 2010-09-19T08:38:39 yup, it's waiting for input (which is good, it hasn't crashed yet) 2010-09-19T08:38:42 to specify when I try to access variables inside planet mono crashes(I think as it times out and it seems to not be running) 2010-09-19T08:38:53 sigh: ok ok xD 2010-09-19T08:39:07 just a second, change to Xchat better ;) 2010-09-19T08:39:12 what hellman said is to open a map file and paste the contents into inthe input 2010-09-19T08:39:13 ok 2010-09-19T08:39:37 *** misato has quit IRC (Quit: Page closed) 2010-09-19T08:40:03 It's not possible for there to be fleets on the very first turn, right? 2010-09-19T08:40:12 *** misato has joined #aichallenge 2010-09-19T08:40:14 re 2010-09-19T08:40:23 javagamer: no 2010-09-19T08:40:42 *** xvinyl has quit IRC (Remote host closed the connection) 2010-09-19T08:40:48 misato: what hellman said is to open a map file and paste the contents into inthe input 2010-09-19T08:40:54 aaah ok 2010-09-19T08:40:57 let's try it 2010-09-19T08:40:58 ;) 2010-09-19T08:40:59 *** xvinyl has joined #aichallenge 2010-09-19T08:41:02 sigh, thanks, just wanted to be sure before I assumed that 2010-09-19T08:41:05 thanks everyone here :) 2010-09-19T08:41:44 sigh, it does nothing too 2010-09-19T08:41:48 don't forget to add "go" line 2010-09-19T08:41:53 oh 2010-09-19T08:42:02 yup, type "go" on the next line, without quotes 2010-09-19T08:42:11 go 2010-09-19T08:42:11 1 11 50 2010-09-19T08:42:12 go 2010-09-19T08:42:28 ... that's your bot giving a command o_O 2010-09-19T08:42:33 bot seems to be ok 2010-09-19T08:42:35 that shouldn't be timing out >< 2010-09-19T08:42:36 i put the firts go 2010-09-19T08:42:43 yes 2010-09-19T08:42:45 i've tried also 2010-09-19T08:43:02 what we did is exactly how the engine talks to your bot 2010-09-19T08:43:06 this 2010-09-19T08:43:06 I fixed this by Using the 'invariant' local, which according to the help files of my development tools was designed for inter-machine communication. 2010-09-19T08:43:15 public static int ParseInt(string value) 2010-09-19T08:43:15 { 2010-09-19T08:43:15 return int.Parse(value, CultureInfo.InvariantCulture); 2010-09-19T08:43:15 } 2010-09-19T08:43:15 public static Double ParseDouble(string value) 2010-09-19T08:43:16 { 2010-09-19T08:43:17 return Double.Parse(value, CultureInfo.InvariantCulture); 2010-09-19T08:43:19 } 2010-09-19T08:43:21 (sorry for the flood) 2010-09-19T08:43:23 *** HakanD has joined #aichallenge 2010-09-19T08:43:40 but when i compile the classes, it shows errors 2010-09-19T08:43:46 sorry, I don't know about that :( 2010-09-19T08:44:14 i've read there's a problem with the locales 2010-09-19T08:44:20 i'm using spanish locale 2010-09-19T08:44:39 yeah, I heard that too, but I'm not using java so I don't know much :( 2010-09-19T08:44:43 xD 2010-09-19T08:44:46 ok thanks anyway 2010-09-19T08:44:47 ;) 2010-09-19T08:44:52 i can try C++ as well 2010-09-19T08:44:53 xD 2010-09-19T08:44:59 :) 2010-09-19T08:45:39 i prefer to use java, but i feel confortable with C++ as well 2010-09-19T08:46:08 by the way, the game is very addictive 2010-09-19T08:46:11 :P 2010-09-19T08:46:16 yes, it is :) 2010-09-19T08:46:38 it's in AppStore as well, isn't it? 2010-09-19T08:46:48 yep 2010-09-19T08:46:53 i'm buying it then xD 2010-09-19T08:46:59 did same (: 2010-09-19T08:47:04 hahaha 2010-09-19T08:47:45 i wonder how this contest effected the sales 2010-09-19T08:48:41 I find it extremly frustrating when you don't know why your bot suddenly stops working in the middle of the game on the server but works well with tcpserver 2010-09-19T08:48:42 it will sell a lot more (i hope) 2010-09-19T08:48:58 sigh, are you coding in c++? 2010-09-19T08:49:03 misato: yes 2010-09-19T08:49:14 how do i launch a test? 2010-09-19T08:49:37 i mean, just to test everything is ok with the example bots 2010-09-19T08:49:39 Is it the original game of that type? I've played Konquest on Linux which is fairly fun 2010-09-19T08:49:51 same way as the java version, but replace "java -jar example_bots/RandomBot.jar" with the path to your executable 2010-09-19T08:49:57 oh ok ^^ 2010-09-19T08:50:25 fahi: I wrote a script to get the playback from the site and feed it to my bot 2010-09-19T08:50:46 sigh: how did you do that? 2010-09-19T08:51:09 someone else did it here: http://www.ai-contest.com/forum/viewtopic.php?f=17&t=628 2010-09-19T08:51:29 reading a1k0n's blog about the tron contest gives lots of good ideas , try it (: 2010-09-19T08:51:33 i implemented mine before I realised this was out there, but same idea 2010-09-19T08:52:34 sigh: i optimize for the highest income for enemies on recapture.... 2010-09-19T08:52:47 HakanD: yeah, I wasn't around for the tron contest, but readind that was interesting 2010-09-19T08:52:48 i thought all the system was in C++ xDD 2010-09-19T08:52:52 i'm stupid 2010-09-19T08:52:53 bartwe: awesome :) 2010-09-19T08:52:59 i've got the same timeouts xD 2010-09-19T08:53:23 misato: looks like a problem with the java engine 2010-09-19T08:53:35 ok, time to try the python engine :P 2010-09-19T08:53:35 oh 2010-09-19T08:53:47 if i pass to the command MyBot 2010-09-19T08:53:52 didn't the java engine have similar problems last time? 2010-09-19T08:54:03 oh yes 2010-09-19T08:54:08 it's a java problem 2010-09-19T08:54:10 c++ has java engine last time i checked 2010-09-19T08:54:12 with the bots 2010-09-19T08:54:18 or has that changed? 2010-09-19T08:54:20 if i use the C++ bot 2010-09-19T08:54:25 it runs well 2010-09-19T08:54:59 sigh, it has java engine 2010-09-19T08:55:07 or i've got java engine here xD 2010-09-19T08:55:15 ok, so c++ bot is working with java engine? 2010-09-19T08:55:21 yep 2010-09-19T08:55:35 i think it's a problem with the java bots 2010-09-19T08:55:41 yeah... 2010-09-19T08:55:42 because the C++ bot runs well 2010-09-19T08:56:00 manzanita:cpp_starter_package misato$ java -jar tools/PlayGame.jar maps/map7.txt 1000 1000 log.txt ./MyBot ./MyBot | java -jar tools/ShowGame.jar 2010-09-19T08:56:05 this, is running well 2010-09-19T08:56:10 Top 10 players: adumlah(3505), kcm1700(3434), McLeopold(3433), Macuyiko(3352), montanalow(3306), paulus(3295), wagstaff(3242), dmj111(3217), golfin(3217), George_Whiteside(3192) 2010-09-19T08:57:00 I just wish I could understand what was wrong with the c# engine,still can't see why it would crash on planet getters. 2010-09-19T08:57:32 misato: well, go with that for now... if you really want java I would try to compile the bots from source 2010-09-19T08:57:46 sigh, i did it as well 2010-09-19T08:57:58 but with no crashes XD 2010-09-19T08:57:59 sigh: thanks, the script works and I do not get any errors replaying the game that just stopped on the server 2010-09-19T08:58:34 fahi: probably a tmeout 2010-09-19T08:59:02 stupids timeouts 2010-09-19T08:59:03 xDD 2010-09-19T08:59:12 and timeouts are wall-clock so it might not even be your fault 2010-09-19T08:59:26 server is under a lot of load from what I hear 2010-09-19T09:00:12 but there is no way to check if it was a timeout, right? 2010-09-19T09:00:23 not that I know of :( 2010-09-19T09:01:12 amstan said new servers are on the way, this would solve this problems for now 2010-09-19T09:01:20 *** phreeza_ has joined #aichallenge 2010-09-19T09:02:38 *** phreeza has quit IRC (Ping timeout: 276 seconds) 2010-09-19T09:03:32 hmmm wtf 2010-09-19T09:03:55 i swear either my distance function or the engine i have isn't working properly 2010-09-19T09:04:15 was the engine changed at all for such errors at any stage? the one that comes with start packages this is 2010-09-19T09:04:21 ah too pesimistic, i assume the enemy can see my new moves in the current move :) 2010-09-19T09:04:56 antimatroid1: I think there have been a few updates 2010-09-19T09:05:11 I know the initial engine had broken battle resolution 2010-09-19T09:06:59 but distance is simple, (int)ceil(sqrt(dx*dx+dy*dy)) so I don't know where that could go wrong (unless it is floating point sneakiness) 2010-09-19T09:07:21 i wonder if some of the timeouts earlier on benzedrine are due to network overload on the server side 2010-09-19T09:08:02 on map 65 my engine thinks the distance between planet 1 and planet 10 is 4 moves, but distance function claims otherwise, i'll go check numbers 2010-09-19T09:08:14 if someone else would be willing to somehow ping that with their engine and look 2010-09-19T09:08:36 I'll have a look 2010-09-19T09:09:22 (int)Math.Ceiling(Math.Sqrt(dx * dx + dy * dy)); 2010-09-19T09:09:31 dunno 2010-09-19T09:09:50 maybe your sqrt returns an int instead of double ? 2010-09-19T09:09:51 I get 4 2010-09-19T09:10:10 at least, my bot thinks the distance is 4 2010-09-19T09:10:24 I'll check what my engine things 2010-09-19T09:10:27 *thiks 2010-09-19T09:10:30 *thinks 2010-09-19T09:10:32 :S 2010-09-19T09:11:56 yeah okay, there's something very wrong with my thing 2010-09-19T09:11:58 *** E9c has joined #aichallenge 2010-09-19T09:12:24 the java engine and my bot both agree on distance 4 2010-09-19T09:12:56 *** Eiii has joined #aichallenge 2010-09-19T09:13:16 when i ping my original distance function it returns 3 2010-09-19T09:13:29 time to have lunch here 2010-09-19T09:13:37 thanks everyone for your help ^^ 2010-09-19T09:14:40 antimatroid1: the euclidean distance I get is 3.20395 2010-09-19T09:14:57 misato: cya 2010-09-19T09:15:04 i suggest your sqrt returns an integer 2010-09-19T09:15:11 thanks sigh i hope to fight against your bot :P 2010-09-19T09:15:17 :) 2010-09-19T09:15:20 see you later 2010-09-19T09:15:21 ^^ 2010-09-19T09:15:46 *** misato has left #aichallenge ("I'll be back!") 2010-09-19T09:15:55 *** boegel has quit IRC (Quit: Leaving) 2010-09-19T09:16:51 *** E9c is now known as agoandanon 2010-09-19T09:16:56 antimatroid1: you declared dx and dx as ints 2010-09-19T09:17:06 sigh: i get that too, trying to work out the error now, nah they're defined as doubles 2010-09-19T09:17:37 you stored the positions as ints 2010-09-19T09:18:33 nope 2010-09-19T09:18:34 hmmm 2010-09-19T09:18:48 damn... psychic debugging fail 2010-09-19T09:19:03 even the constructor is claiming double parameters properly 2010-09-19T09:19:07 sigh: raymondchen++ 2010-09-19T09:19:08 oh wait 2010-09-19T09:19:40 you declared *something* as an int 2010-09-19T09:19:43 (: 2010-09-19T09:19:54 ahuh, in my parseGameStateSetup I added variables today to simplify things and did those as ints 2010-09-19T09:19:55 thanks 2010-09-19T09:20:00 saved me hours of frustration 2010-09-19T09:20:16 man, I should have known that to be a psychic you need otbe as vague as possible 2010-09-19T09:20:47 ha, seriously :P 2010-09-19T09:20:50 :D 2010-09-19T09:21:18 so I was right the second time, positions were stored as ints :P 2010-09-19T09:22:05 takes forever...... http://www.benzedrine.cx/planetwars/canvas?game_id=42748 2010-09-19T09:23:21 draw! http://www.benzedrine.cx/planetwars/canvas?game_id=42250 2010-09-19T09:23:28 slightly non-trivial 2010-09-19T09:24:02 sigh: you got him on the ropes 2010-09-19T09:24:08 sigh: sort of, only when i pulled them on the first move from the game state string crap 2010-09-19T09:24:22 *** jaspervdj has joined #aichallenge 2010-09-19T09:24:23 and that is buried away in a function that i hack stuff into 2010-09-19T09:24:41 sigh: i would judge this one as a win for you 2010-09-19T09:24:44 so it would have taken forever to find the bug if you hadn't said that 2010-09-19T09:25:00 bartwe: why? 2010-09-19T09:25:01 *** Appleman1234 has quit IRC (Ping timeout: 252 seconds) 2010-09-19T09:25:04 dhartmei makes little threat 2010-09-19T09:25:23 but we are still dead even 2010-09-19T09:25:28 both of you could have taken the 89 neutral and broken it 2010-09-19T09:25:58 sigh: my instincts say that you have him in defense, giving you more influence 2010-09-19T09:26:07 my bot seems to have done slightly better recently on the official site 2010-09-19T09:26:14 hoping my new one will be way better than that 2010-09-19T09:26:25 bartwe: results are what counts :P 2010-09-19T09:26:25 he fails to use his advantage in numbers due to your travel times to take a neutral 2010-09-19T09:27:12 hmm... I'm not as certain as you are that taking that neutral is good 2010-09-19T09:27:36 might have a play around with this map and see if you are right :) 2010-09-19T09:28:19 you need a neutral to break the stalemate in numbers, that or superior tactics, but even that would just mean gaining better growthrate 2010-09-19T09:30:02 *** Utkarsh has quit IRC (Ping timeout: 264 seconds) 2010-09-19T09:30:56 *** sannysanoff has joined #aichallenge 2010-09-19T09:31:17 any orgs over here? 2010-09-19T09:32:06 orgs? 2010-09-19T09:32:12 orgz 2010-09-19T09:32:18 contest organizers 2010-09-19T09:32:41 if they will read the irc log, please, orgs, add "vector" package (cabal install vector) 2010-09-19T09:32:44 *** Eiii has quit IRC (Ping timeout: 252 seconds) 2010-09-19T09:32:52 for haskell 2010-09-19T09:33:03 you are better off asking in the forums, probably 2010-09-19T09:33:16 i am going in that direction ;-) 2010-09-19T09:33:21 *** mega1 has joined #aichallenge 2010-09-19T09:33:23 :) 2010-09-19T09:33:35 or open an issue (: 2010-09-19T09:33:57 http://code.google.com/p/ai-contest/issues/list 2010-09-19T09:34:25 *** iFire has quit IRC (Remote host closed the connection) 2010-09-19T09:34:56 *** iFire has joined #aichallenge 2010-09-19T09:35:00 @rankings 2010-09-19T09:35:01 sigh: Top 10 players: adumlah(3488), kcm1700(3463), McLeopold(3435), paulus(3351), montanalow(3304), sighbot(3275), Macuyiko(3259), wagstaff(3232), golfin(3215), dmj111(3206) 2010-09-19T09:35:08 awesome! back in the top 10 2010-09-19T09:35:30 and I got #1 and #3 on tcp 2010-09-19T09:35:35 today's been a good day :) 2010-09-19T09:36:06 *** Utkarsh has joined #aichallenge 2010-09-19T09:36:06 gratz! 2010-09-19T09:36:18 thank you! 2010-09-19T09:36:38 gz, keep it up (: 2010-09-19T09:37:41 and i'm slipping away out of the top10 again 2010-09-19T09:38:22 don't worry, it was just to make a place for me :P 2010-09-19T09:38:26 hehe 2010-09-19T09:39:08 still a long time to go before the end of the contest though, I'm wondering if people will still come up with some radical changes in strategy 2010-09-19T09:39:18 sure 2010-09-19T09:39:51 someone will figure out how to break the branching problem and start doing minimax 2010-09-19T09:40:03 ugh, then I'm boned 2010-09-19T09:40:06 or somethign similair 2010-09-19T09:40:42 how do you apply minimax when the turns are interleved? 2010-09-19T09:41:26 * sigh goes and cracks open his AI book 2010-09-19T09:41:41 *not interleved 2010-09-19T09:42:07 interleaved ? 2010-09-19T09:42:07 good question 2010-09-19T09:42:11 not sure what you mean by that 2010-09-19T09:42:38 ah, now i do 2010-09-19T09:42:49 let me rephrase: how do you apply minimax when the turns are simultaneous 2010-09-19T09:43:08 you don't 2010-09-19T09:43:33 perfect mirror play starts to come to mind 2010-09-19T09:43:43 which is useless for searching 2010-09-19T09:46:52 wasn't it the same problem for tron? 2010-09-19T09:47:09 tron was simultaneous as well? 2010-09-19T09:47:19 yes, and we still used minimax 2010-09-19T09:47:30 how does that work? 2010-09-19T09:47:46 nobody discovered the better way :) 2010-09-19T09:47:47 assume that your opponent knows your move? 2010-09-19T09:47:52 yes 2010-09-19T09:48:02 that's what I think is more pessimistic 2010-09-19T09:48:05 ok, that makes sense 2010-09-19T09:48:09 thank you 2010-09-19T09:48:22 assuming that you know your oppenent's move is rather optimistic 2010-09-19T09:48:24 sigh: what if you simply assume the opponent does nothing ? 2010-09-19T09:48:26 * Naktibalda should read what the minimax is :) 2010-09-19T09:48:49 *** user42 has joined #aichallenge 2010-09-19T09:48:53 bartwe: i'm not following you 2010-09-19T09:49:19 sigh: how powerfull would minimax on this be if you just assume the enemy is passive other then the allready made movements ? 2010-09-19T09:50:33 it might be a good heuristic, but my instincts say no (my instincts have failed me many times already) 2010-09-19T09:50:43 bartwe: where is the min phase in that? 2010-09-19T09:50:55 hi from cologne, germany. anybody can help with jar-ing MyBot to use it like the sample bots ? i allways get an exception after "jar -cf *.class" and using the Bot with "java -jar MayBot.jar" 2010-09-19T09:51:39 mega1: make a null phase 2010-09-19T09:51:57 oh, a solitare 2010-09-19T09:51:59 single transisition that does nothing as the min 2010-09-19T09:52:16 user42: what's the exception? 2010-09-19T09:52:39 sigh: Exception in main: NullPointer 2010-09-19T09:52:43 you don't do any game tree search yet? 2010-09-19T09:53:15 sigh: and in Thread-2: ArrayIndexOutOfBound 2010-09-19T09:53:40 sounds like a bug in your bot 2010-09-19T09:53:53 or are you just compiling the example? 2010-09-19T09:55:00 Is there any way to tell what turn you're on from the data you get from the java client? 2010-09-19T09:55:11 no 2010-09-19T09:55:15 sigh: the bot is running fine when running with "java MyBot.java" but not after "jar -cf MyBot.jar *.class" and Running it with "java -jar Mybot.jar" 2010-09-19T09:55:30 Naith: just keep a counter in the main loop 2010-09-19T09:55:31 *** pc has joined #aichallenge 2010-09-19T09:55:49 Naith: the data isn't given explicitly 2010-09-19T09:56:01 ooo okay thanks! 2010-09-19T09:56:10 Top 10 players: kcm1700(3449), McLeopold(3448), adumlah(3430), paulus(3337), montanalow(3292), sighbot(3267), wagstaff(3217), golfin(3208), Macuyiko(3203), dmj111(3196) 2010-09-19T09:56:12 *** xvinyl has quit IRC (Remote host closed the connection) 2010-09-19T09:56:26 user42: then sorry, I don't know... I'm not that familiar with java 2010-09-19T09:56:36 sigh: ok, thanks 2010-09-19T09:56:59 Anyone here, who is programming in Java ? 2010-09-19T09:57:55 *** antimatroid has joined #aichallenge 2010-09-19T09:58:05 yea 2010-09-19T09:58:13 *** antimatroid1 has quit IRC (Ping timeout: 245 seconds) 2010-09-19T09:58:34 delt0r___: u tried to make an jar form your robot ? 2010-09-19T09:58:48 haven't needed to 2010-09-19T09:59:04 but it would be easy to do 2010-09-19T09:59:14 delt0r___: running different versions of your bot against each other ? 2010-09-19T09:59:20 delt0r___: ok, how ? :-) 2010-09-19T09:59:42 well its just a normal jar... with a manifiest file... 2010-09-19T09:59:44 jar --help 2010-09-19T09:59:54 delt0r___: i tried "jar -cf MyBot.jar *.class" 2010-09-19T09:59:58 or use the IDE you are using to do it for you 2010-09-19T10:00:06 or an ant task 2010-09-19T10:00:26 you need to tell it what class is the "main" 2010-09-19T10:00:45 so that java -jar MyBot.jar 2010-09-19T10:00:48 will work 2010-09-19T10:00:53 otherwise just use 2010-09-19T10:01:03 java -cp MyBot.jar MyBot 2010-09-19T10:01:03 delt0r___: ok, thank u very much. i'll check this 2010-09-19T10:01:28 delt0r___: java -cp ... in stead of jar -cf ? 2010-09-19T10:01:41 no to "run" the jar 2010-09-19T10:01:49 delt0r___: ok, thanks 2010-09-19T10:01:50 so for your bot command line 2010-09-19T10:03:11 *** phreeza_ has quit IRC (Quit: Computer has gone to sleep) 2010-09-19T10:04:00 ouch: http://www.benzedrine.cx/planetwars/getplayer?player=sigh-1 2010-09-19T10:04:09 that's a signal for me to stop 2010-09-19T10:04:16 delt0r___: ok, works perfectly. thanks 2010-09-19T10:04:22 np 2010-09-19T10:05:29 Hello 2010-09-19T10:05:36 do you have an idea of how to kill ragebot ? 2010-09-19T10:05:37 delt0r___: i read some ppost in the forum about an python lib with functions for simulating the "future". is there some for java? i've done some simple simulation in MyBot 2010-09-19T10:06:02 user42: I don't know... i write all my own code. 2010-09-19T10:06:03 user42: I wrote a recursive method to see the future of the game, but it doesn't help so much 2010-09-19T10:06:12 elijahbal: work on your defence 2010-09-19T10:06:32 I have a java player... but for the game "model" i use a simpler model 2010-09-19T10:06:58 I see the future based only on existing fleets + fleets I plan to launch this turn 2010-09-19T10:07:04 Defense .. Yes I do not have a « real defense » but for map18.txt, you can see that rage bot is really near of my planet so it doesn't let me so much time. 2010-09-19T10:08:00 elijahbal: I wouldn't worry about those maps with close planets right away... make sure you can beat ragebot on most other maps 2010-09-19T10:08:00 so, if enemy is close to you, apply your defense strategy accordingly 2010-09-19T10:08:31 but yes, close enemy requires more sophisticated defence 2010-09-19T10:08:36 top bots gives some good ideas about this (: 2010-09-19T10:08:38 delt0r___: which is your bot ? 2010-09-19T10:08:38 *** javagamer has quit IRC (Ping timeout: 245 seconds) 2010-09-19T10:09:13 elijahbal: in map18, bots start in oposite corners 2010-09-19T10:09:42 user42: Its not on the main site yet... I am still debuging my data structs 2010-09-19T10:09:57 Naktibalda: I meant map30 2010-09-19T10:10:12 hey, oh the java game engine (1.2) is buggy 2010-09-19T10:10:22 right now i am finsihing my java game player, TCP watcher and humanBorg player 2010-09-19T10:10:28 then i will release it 2010-09-19T10:10:43 delt0r___: ok. cool. her is mine: http://ai-contest.com/profile.php?user_id=8315 2010-09-19T10:10:45 Draw 2010-09-19T10:11:10 user42: not bad 2010-09-19T10:11:16 probably I get Draws in 4 or 6 maps where my bot can't win 2010-09-19T10:12:04 my bot should take a planet in a corner, probably its' too far 2010-09-19T10:13:03 Naktibalda: A draw? that would be hard wouln't it... i mean to have the same number of ships at turn 200? 2010-09-19T10:13:07 one small adjustment and my bot wins 2010-09-19T10:13:50 it remainded 1:1 situation, because my bot didn't dare to capture a planet in a corner 2010-09-19T10:13:59 I had max_distance/2 limit for attacks 2010-09-19T10:15:34 http://pastebin.ca/1944178 <-- look 2010-09-19T10:16:44 hmm, does anybody work on pastebin for visualizer? If not, I could try to do it 2010-09-19T10:16:48 *** whereby has joined #aichallenge 2010-09-19T10:17:27 *** jmreardon has joined #aichallenge 2010-09-19T10:18:47 elijahbal: does it have to end at turn 13? 2010-09-19T10:20:44 no 2010-09-19T10:20:59 but the game engine stop the game and says me Player 2 lose 2010-09-19T10:21:10 at turn 13? or later 2010-09-19T10:21:48 no the game doesn't continue 2010-09-19T10:21:53 but it should I think 2010-09-19T10:23:31 your bot crashed or did an invalid move 2010-09-19T10:24:53 what a frustrating loss: http://www.ai-contest.com/visualizer.php?game_id=4581708 >< 2010-09-19T10:26:24 sigh got very close to losing 2010-09-19T10:26:29 good defense (: 2010-09-19T10:26:44 the single tick loss of the one planet would shift it away from him 2010-09-19T10:26:58 bartwe: sigh did lose 2010-09-19T10:27:17 ah, point in case 2010-09-19T10:27:17 by 4 planets 2010-09-19T10:27:21 move 108 2010-09-19T10:27:24 single tick loss of parity 2010-09-19T10:27:56 I had to parse the playback string to figure that one out 2010-09-19T10:28:23 defence definitely needs work 2010-09-19T10:28:36 my guess is that these stalemates will become much more common with the top bots 2010-09-19T10:29:01 sigh: I am surprised... 2010-09-19T10:29:21 Its not a very good AI if it can't work out that its not getting ahead 2010-09-19T10:29:40 my AI doesn't know anything about who's ahead :P 2010-09-19T10:29:40 and after 200 turns with ships==ships 2010-09-19T10:29:52 it doesn't know the ship count or the total growth rate :P 2010-09-19T10:29:55 oh... Mine does 2010-09-19T10:29:56 quite stupid :D 2010-09-19T10:29:59 income and ships 2010-09-19T10:29:59 delt0r: mirror play is nearly unavoidable for the best ais 2010-09-19T10:30:07 NewBot vs RageBot 99/100 Avg Turns: 78.0 2010-09-19T10:30:16 elijahbal: thanks for a map :) 2010-09-19T10:30:40 hmm, I have to modify test.py to find that strange map where my bot can't win 2010-09-19T10:30:43 bartwe: I don't really think so... Unless your getting really close to the "optimal" aka minmax search 2010-09-19T10:31:00 delt0r: optimal is the goal here 2010-09-19T10:31:10 didn't Zannick have copybot on the tcp? 2010-09-19T10:31:14 but impractical 2010-09-19T10:31:18 I recall it didn't do that well 2010-09-19T10:31:34 sigh: its has to be one turn out of sync 2010-09-19T10:31:50 you don't get to see what the other is doing till after your done 2010-09-19T10:31:57 delt0r___: yes, I know 2010-09-19T10:32:14 so thats an automatic copybot==looser 2010-09-19T10:32:28 try copying moves in chess 2010-09-19T10:33:14 this is more symmetrical than chess, and plenty of places with 2nd move advantage 2010-09-19T10:34:19 there is no second move advantage if you are "constrained" to make the same move 2010-09-19T10:34:33 because you use none of the extra information 2010-09-19T10:35:13 I guess I'll have to code one up myself, it will be good for testing anyway 2010-09-19T10:35:25 right now is sleep though 2010-09-19T10:35:28 good night 2010-09-19T10:35:37 it was fun talking to you all 2010-09-19T10:35:38 l8r 2010-09-19T10:35:58 *** sigh has quit IRC (Quit: Leaving...) 2010-09-19T10:38:37 bartwe: We can always solve that problem (mirror games) with real random maps ;) 2010-09-19T10:39:12 *** iioo has joined #aichallenge 2010-09-19T10:42:04 oh, in map26 rage bot manage to capture my home planet :( 2010-09-19T10:43:01 I should check a distance to closest enemy planet 2010-09-19T10:43:07 delt0r: also i mean true copyplay, so future predicting in this concurrent move game 2010-09-19T10:46:44 *** ntroPi has quit IRC () 2010-09-19T10:47:29 delt0r: real random maps would give unfair advantages 2010-09-19T10:47:48 would require larger sample sizes for representative outcomes 2010-09-19T10:47:56 also 'real' random is hard to get 2010-09-19T10:48:28 bartwe: Yea.. you would need a very large number of games --since i think even a bad bot would win quite often in random maps 2010-09-19T10:50:03 I am not actually sugesting random maps btw.. I still favour rules that permit large numbers of games 2010-09-19T10:50:18 argh 2010-09-19T10:50:26 i've been throwing away perfectly good bots 2010-09-19T10:50:35 ecause i infact wasn't testing them 2010-09-19T10:50:43 *** Blkt has joined #aichallenge 2010-09-19T10:56:10 Top 10 players: kcm1700(3468), McLeopold(3453), adumlah(3451), paulus(3359), montanalow(3317), wagstaff(3272), golfin(3268), sighbot(3262), dmj111(3211), Macuyiko(3201) 2010-09-19T10:57:11 bartwe: bartwe.10 is losing on invalid input right off the bat on tcp 2010-09-19T10:57:29 yeah i noticed 2010-09-19T10:57:32 hmz 2010-09-19T10:57:37 time for the debug tools 2010-09-19T11:07:48 anyone have a script to run your bot against all maps/defaultbots and give ome statistics of duration/win/lose ? 2010-09-19T11:08:40 bartwe: I have a java one... its built into my player... 2010-09-19T11:08:53 ah 2010-09-19T11:08:55 that works 2010-09-19T11:09:26 well yea... its desinged to support not restarting bots for the process setup and teardown 2010-09-19T11:09:47 since thats pretty slow with java... but that would work with anything... 2010-09-19T11:10:22 the bot just needs to reset isself with a "reset" --or just use default behavour 2010-09-19T11:10:44 *** pc has quit IRC (Remote host closed the connection) 2010-09-19T11:10:56 but i need at least an hour before i am finished all the other stuff... did you want to use it? 2010-09-19T11:11:12 *** pc has joined #aichallenge 2010-09-19T11:11:22 nah 2010-09-19T11:11:27 my bot isn't in java 2010-09-19T11:11:34 I know 2010-09-19T11:11:48 and its not hard to this all in the shell either 2010-09-19T11:12:15 my new bots are utter crap 2010-09-19T11:13:05 the more complex my bot gets--the worse it does 2010-09-19T11:13:09 hehe 2010-09-19T11:13:33 thats why i tried simplifying 2010-09-19T11:13:47 so far no luck 2010-09-19T11:15:23 *** Fumpful has joined #aichallenge 2010-09-19T11:16:11 *** espes has quit IRC (Read error: Connection timed out) 2010-09-19T11:17:25 *** espes has joined #aichallenge 2010-09-19T11:25:54 *** naglijo has joined #aichallenge 2010-09-19T11:28:19 *** sinan has joined #aichallenge 2010-09-19T11:32:06 *** nood has joined #aichallenge 2010-09-19T11:33:20 *** HakanD has quit IRC (Quit: Leaving) 2010-09-19T11:33:41 *** HakanD has joined #aichallenge 2010-09-19T11:41:38 *** delt0r_ has joined #aichallenge 2010-09-19T11:43:46 *** delt0r___ has quit IRC (Ping timeout: 272 seconds) 2010-09-19T11:44:01 *** amstan has joined #aichallenge 2010-09-19T11:44:01 *** ChanServ sets mode: +o amstan 2010-09-19T11:44:36 *** pjreddie has quit IRC (Quit: Page closed) 2010-09-19T11:44:55 *** ermau has joined #aichallenge 2010-09-19T11:45:07 *** Naith has quit IRC (Quit: Naith) 2010-09-19T11:46:22 *** bourbaki has joined #aichallenge 2010-09-19T11:46:45 greetings 2010-09-19T11:46:51 salutations 2010-09-19T11:46:57 alright new bots going on tcp now 2010-09-19T11:47:15 do you discuss strategies in here? 2010-09-19T11:47:24 i tend not to 2010-09-19T11:47:24 why not 2010-09-19T11:47:27 you are welcome to 2010-09-19T11:47:34 only if you want your enemy to be stronger :) 2010-09-19T11:47:47 sharing is caring... 2010-09-19T11:47:57 im not participating just would like to talk about some strategies 2010-09-19T11:48:23 has anyone tried out ant colony optimization yet? 2010-09-19T11:48:33 bourbaki: Epic fail 2010-09-19T11:48:43 heh 2010-09-19T11:48:49 *** frantik has joined #aichallenge 2010-09-19T11:48:51 so yes i did... but i did expect it to fail 2010-09-19T11:48:59 so the better bots are all handcrafted? 2010-09-19T11:49:20 is there any way to train the bots offline? 2010-09-19T11:49:29 It may work to decide "what to do" but not in game... in fact i doubt it would work then either 2010-09-19T11:49:36 sure 2010-09-19T11:49:39 *** sugarshark has joined #aichallenge 2010-09-19T11:50:11 but you need to "preserve" the offline/online bit for the server... you are not alwoed to write to disk 2010-09-19T11:50:14 well what i thought was that id like to model all the support functions in the ant colony in a vector, then map that with a SOM 2010-09-19T11:50:26 and your bot is restarted each time 2010-09-19T11:50:47 to have some strategies and reverse engineer the problem space with the viterbi algorith (HMM) to categorize the enemy 2010-09-19T11:50:59 interesting... 2010-09-19T11:51:08 yes but i could compile the strategy as the source code 2010-09-19T11:51:20 so id train the ai offline agains the SOM categories 2010-09-19T11:51:31 *** Kastra has quit IRC (Ping timeout: 240 seconds) 2010-09-19T11:51:34 Whats the hidden states in this case? 2010-09-19T11:51:35 then use the HMM to determine the other guys strategy 2010-09-19T11:51:41 oh ic 2010-09-19T11:51:42 lets see how this goes 2010-09-19T11:51:52 hidden states would be the support functions the probability etc 2010-09-19T11:52:12 and then use the online optimization to adapt 2010-09-19T11:52:22 timeout against ragebot, great start 2010-09-19T11:53:20 how does a random bot rank? 2010-09-19T11:53:47 bourbaki: badly 2010-09-19T11:53:52 very badly 2010-09-19T11:53:54 bourbaki: probably tight after the stallbots 2010-09-19T11:54:05 another thing that would interest me is if there could be some kind of proof if there is a single best or not loose strategy 2010-09-19T11:54:28 bourbaki: I discussed that with a friend 2010-09-19T11:54:45 bourbaki: two perfect bots would be locked in a pure mirror stalemate i think 2010-09-19T11:54:46 i thought that a petri net might be an idea to start a proof 2010-09-19T11:54:49 bourbaki: if the objective was to lose, it would actually be a hard problem too 2010-09-19T11:54:50 or lattice bolzmann 2010-09-19T11:54:52 bourbaki: exsitance of such a thing should be provable... however i don't think its in any way practical 2010-09-19T11:55:32 well i think its totally useless to handcraft something if you can search the space for a strategy 2010-09-19T11:55:43 very meta 2010-09-19T11:55:49 searching for a search strategy 2010-09-19T11:55:52 *** blbrown_win3_ has quit IRC (Quit: Leaving) 2010-09-19T11:55:52 but if you know that there is a single best strategy then the search is a lot different 2010-09-19T11:55:56 bourbaki: true --but the state space is much larger than tron 2010-09-19T11:56:09 *** frantik has quit IRC (Quit: Page closed) 2010-09-19T11:56:10 Top 10 players: kcm1700(3462), McLeopold(3460), adumlah(3448), paulus(3373), golfin(3318), montanalow(3310), wagstaff(3271), EBraun(3232), sighbot(3221), Longing(3192) 2010-09-19T11:56:17 bartwe as i said earlier eg, the vector for the aco algorithm 2010-09-19T11:56:20 *** syntaxglitch has joined #aichallenge 2010-09-19T11:56:24 its not go... but is closer to go than chess 2010-09-19T11:56:29 delt0r heh true 2010-09-19T11:57:08 did anyone try to decrease the problem space, for a model? 2010-09-19T11:57:15 there exists an optimal strategy 2010-09-19T11:57:19 i know very fuzzy 2010-09-19T11:57:21 it's a draw 2010-09-19T11:57:39 how does the ranking work? would always draw get you to top? 2010-09-19T11:57:41 assuming both players play it 2010-09-19T11:57:59 it's some kind of elo ranking system, not sure if that would work or not 2010-09-19T11:58:03 bourbake: depends, if you draw highly ranked players you do 2010-09-19T11:58:10 if the bots are smart theyd solve the prisioners problem and noone attacks each other ;0 ? 2010-09-19T11:58:11 but likely you wont get selected to play them 2010-09-19T11:58:12 you shouldn't be drawing against shit players 2010-09-19T11:58:27 bourbaki: it's not always that simple 2010-09-19T11:59:00 are you guys working on a analysis toolbox? 2010-09-19T11:59:10 then to shake things up... push the map sizes up a bit... or have 3 way games. 2010-09-19T11:59:32 bourbaki: so your thinking of an academic approach ? 2010-09-19T11:59:40 bartweyes 2010-09-19T11:59:41 now your state space is really large and optimal play will not be possible with 1 sec turn time 2010-09-19T12:00:00 not interested in the real time problem really but in getting more into the problem and its model 2010-09-19T12:00:14 delt0r agree 2010-09-19T12:00:19 it's a lot harder to try and implement an optimal solution than you might think 2010-09-19T12:00:31 bourbaki: the concurrent move side of the proble is stumping the usual approaches 2010-09-19T12:00:34 suggestions ? 2010-09-19T12:00:35 but suppose you could decrease the problems dimensions by the analysis 2010-09-19T12:01:01 bartwe pardon? 2010-09-19T12:01:15 bartwe: In Academia you change the problem... make it round robin ;) 2010-09-19T12:01:39 you can decrease it, but not that much 2010-09-19T12:01:46 i love game theory, but it's really not all that useful for this 2010-09-19T12:01:53 hehe, as i said im more interested in a model, the math behind it 2010-09-19T12:01:55 delt0r: feel free to make projections, the core description doesn't change 2010-09-19T12:02:46 bourbaki: luckily the problem has resisted such anaylis 2010-09-19T12:03:09 or someone isn't sharing 2010-09-19T12:03:16 has it been a big problem already before the contest? 2010-09-19T12:03:29 not one that i recognize 2010-09-19T12:03:38 bartwe: Of course 2010-09-19T12:03:38 i guess it could be related to poker or stockmarkets 2010-09-19T12:04:20 do you have network access in the program? 2010-09-19T12:04:21 http://www.benzedrine.cx/planetwars/canvas?game_id=44581 2010-09-19T12:04:24 talk about retarded aha 2010-09-19T12:04:26 anyone else getting a lot of timeouts on tcp? 2010-09-19T12:04:31 i am 2010-09-19T12:04:35 bartwe: The fleets and the fact thay take a while changes the problem quite a bit 2010-09-19T12:04:39 antimatroid: benz isn't working for me 2010-09-19T12:04:53 *** sinan has quit IRC (Ping timeout: 245 seconds) 2010-09-19T12:08:49 bartwe: works fine here :\ 2010-09-19T12:09:50 danielvf, dstufft: do you guys know how to check how many concurrent games we're running 2010-09-19T12:12:20 *** atmz443 has quit IRC (Quit: atmz443) 2010-09-19T12:17:33 gah, so annoying when i accidentally close tcp window mid match 2010-09-19T12:19:46 *** bradtgmurray has quit IRC (Remote host closed the connection) 2010-09-19T12:20:20 antimatroid: hmz, i might be blocked then 2010-09-19T12:20:45 bots are still running 2010-09-19T12:20:50 but the website is not 2010-09-19T12:21:10 *** bradtgmuray has joined #aichallenge 2010-09-19T12:21:39 don't know why you'd get blocked, or having you been doing something shady? :P 2010-09-19T12:21:47 nope 2010-09-19T12:21:53 no crawling or anything 2010-09-19T12:22:25 Is there a newer version of PlayGame out that's newer than PlayGame-1.2.jar? 2010-09-19T12:22:41 antimatroid: works from another host of mine, great... 2010-09-19T12:22:46 not that i know of. there is some c++ engine on the forums i haven't tried though 2010-09-19T12:23:17 bradtgmuray: the repository might have a newer version of PlayGame.jar, but i dunno for sure 2010-09-19T12:23:31 antimatroid: also works on another machine in the local network, ok this is intresting :) 2010-09-19T12:23:35 Hrmf 2010-09-19T12:23:54 I'm having issues getting stderr out to the console with it with a python bot 2010-09-19T12:24:33 bradtgmuray: http://github.com/ai-contest/planetwars-viz i think that's where PlayGame comes from 2010-09-19T12:24:52 there is a makefile target for it 2010-09-19T12:25:06 *** sinan has joined #aichallenge 2010-09-19T12:25:14 jmcarthur: I've also seen a link to a repo on google code for the same stuff 2010-09-19T12:25:46 bradtgmuray: we migrated to github 2010-09-19T12:25:54 hmmm, doesn't seem like anything special so far 2010-09-19T12:26:05 wait what? 2010-09-19T12:26:26 ? 2010-09-19T12:26:36 the ai-contest code is on github now? 2010-09-19T12:26:40 yeah 2010-09-19T12:26:51 hasn't it been like this this entire contest? 2010-09-19T12:26:54 it was kind of quiet though 2010-09-19T12:27:03 i thought everybody was informed (i didn't do it) 2010-09-19T12:27:13 hey, I have a quick hack here: http://visualizer.naktibalda.lt/2 what do you think about it? 2010-09-19T12:28:49 fpor the visualizer 2010-09-19T12:29:06 keep track of fleet countts,growth, grounded etc. 2010-09-19T12:29:16 a1k0n: dstufft did it. originally he was going to have jeff take the svn repo down, but we figured it might be better to leave it up for a few more days in case somebody had changes in working copies to finish first. apparently we also need to make sure everybody knows 2010-09-19T12:29:21 have a bar on the timeline 2010-09-19T12:29:25 for the current moment 2010-09-19T12:29:32 allow clicking on the timeline 2010-09-19T12:29:43 allow replay speed to be changed. 2010-09-19T12:29:56 heck, i want to be able to put annotations on the timeline too :) 2010-09-19T12:30:39 86. 1-3x57!? 2010-09-19T12:31:14 wat 2010-09-19T12:31:32 Naktibalda: oh that's rad 2010-09-19T12:32:15 rad? 2010-09-19T12:32:31 is it about my design skills? 2010-09-19T12:32:48 heh. no, i mean the fact that you can upload games to share them. i wanted that yesterday. 2010-09-19T12:33:56 *** HakanD has quit IRC (Ping timeout: 240 seconds) 2010-09-19T12:39:01 *** HakanD has joined #aichallenge 2010-09-19T12:39:12 dhartmei_: hi 2010-09-19T12:39:13 There we go 2010-09-19T12:39:24 If you don't print newlines on your stderr messages they don't come out 2010-09-19T12:41:11 after a week, now im getting somewhere 2010-09-19T12:41:26 bradtgmuray: use the modified tournament engine, they do come out in that one 2010-09-19T12:42:22 Anybody could tell me why I use the TCP to test my bot. and always get the time out error? And in the server 2010-09-19T12:43:30 whereby: i have this problem when my ping time to the TCP server is very high 2010-09-19T12:44:06 thx. maybe I have the same problem... 2010-09-19T12:44:52 *** Blkt has quit IRC (Ping timeout: 265 seconds) 2010-09-19T12:45:00 whereby: you can try to ping the server :) 2010-09-19T12:46:31 whereby: do you get timeouts in official server? 2010-09-19T12:47:03 yes. 2010-09-19T12:48:20 ls 2010-09-19T12:48:25 sorrty 2010-09-19T12:48:26 here is the video for the time out http://www.benzedrine.cx/planetwars/canvas?game_id=44977 2010-09-19T12:49:12 whereby: if you are getting timeouts on the official server too then your code is not fast enough :) 2010-09-19T12:49:46 whereby: you are allocated a particular time to finish one move, your code has to finish the turn by that time 2010-09-19T12:50:10 It is not on the offical server, on the offical server ,my code could run withou error 2010-09-19T12:51:05 whereby: i see. Naktibalda just asked you above if you get timeoutes on the official server, and I thought you said "yes" :) 2010-09-19T12:51:50 :) 2010-09-19T12:53:36 So, bad luck ,I could only wait for the offical server result and watch the result.:) 2010-09-19T12:56:10 Top 10 players: McLeopold(3459), kcm1700(3446), adumlah(3390), golfin(3375), paulus(3372), montanalow(3295), EBraun(3286), wagstaff(3256), Fyodor(3212), sighbot(3207) 2010-09-19T12:57:39 *** smf68 has quit IRC (Remote host closed the connection) 2010-09-19T12:57:43 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-19T12:57:58 *** danielvf has joined #aichallenge 2010-09-19T12:59:30 *** whereby has left #aichallenge 2010-09-19T13:02:28 finally took the trouble to import tcp.c into my own code 2010-09-19T13:02:35 should help with some debugging 2010-09-19T13:02:58 ok, i'm moving the tcp server to an amazon ec3 instance in the us, should solve those timeout problems and lag. 2010-09-19T13:03:05 still testing, i'll switch in about 30 minutes 2010-09-19T13:03:24 dhartmei_: thanks for setting a bed time 2010-09-19T13:03:39 my lecturers would thank you 2010-09-19T13:03:53 heh, there will only be downtime of a couple of seconds ;) 2010-09-19T13:04:48 dhartmei_: you rock 2010-09-19T13:05:08 now i only need an ec node for my end 2010-09-19T13:05:29 bartwe: you could the new micro nodes, only around $14 per month :) 2010-09-19T13:05:44 dhartmei_: also, does your firewall autoblock ? 2010-09-19T13:08:22 it did, but i disabled that a few days ago 2010-09-19T13:09:13 i'll sync the latest couple of hundred most recent games, and the old ones later, it's several gigs, and i can't transfer them fast (my uplink). 2010-09-19T13:15:40 *** pc has quit IRC (Remote host closed the connection) 2010-09-19T13:16:03 *** pc has joined #aichallenge 2010-09-19T13:17:00 are planets with larger growthrate bigger? 2010-09-19T13:17:13 fahi: yes 2010-09-19T13:17:32 antimatroid: alright, ty 2010-09-19T13:21:05 *** tobiassjosten has joined #aichallenge 2010-09-19T13:21:48 *** rwa has quit IRC (Quit: rwa) 2010-09-19T13:31:08 *** sinan has quit IRC (Ping timeout: 240 seconds) 2010-09-19T13:33:57 *** Eiii has joined #aichallenge 2010-09-19T13:38:44 *** dhartmei has joined #aichallenge 2010-09-19T13:39:06 *** naglijo has quit IRC (Ping timeout: 255 seconds) 2010-09-19T13:39:49 Since I registered for the contest my login for the forums (that lived on since the tron times) stopped working. 2010-09-19T13:40:20 And 'forgot password' sends no mail. 2010-09-19T13:40:23 amstan? 2010-09-19T13:44:04 *** Theobon has joined #aichallenge 2010-09-19T13:45:40 Any idea when we can change our bots name, or delete accounts from contest? 2010-09-19T13:46:28 *** mewo2 has joined #aichallenge 2010-09-19T13:49:52 tcp server moved: ip 72.44.46.68 port 995 2010-09-19T13:53:32 *** HakanD has quit IRC (Ping timeout: 240 seconds) 2010-09-19T13:53:43 dhartmei: <3 2010-09-19T13:54:05 *** Eiii has quit IRC (Ping timeout: 272 seconds) 2010-09-19T13:55:55 dhartmei: cool 2010-09-19T13:56:02 Seems a lot faster now. 2010-09-19T13:56:09 dhartmei: am i firewalled on the website btw ? can't connect 2010-09-19T13:56:10 Top 10 players: kcm1700(3500), McLeopold(3468), adumlah(3382), golfin(3367), paulus(3346), EBraun(3319), montanalow(3316), Fyodor(3268), wagstaff(3206), dmj111(3202) 2010-09-19T14:01:04 the website is http://72.44.46.68/ 2010-09-19T14:01:23 you can't access that one? then amazon is firewalling you... :) 2010-09-19T14:02:23 oho 2010-09-19T14:03:58 cool, you even have an error message for that 2010-09-19T14:05:22 wow this is much faster :D 2010-09-19T14:05:50 * antimatroid is not looking forward to people using 1 second per turn 2010-09-19T14:11:44 i can't even keep up with this number of games 2010-09-19T14:14:56 mega1: i need your username and email address 2010-09-19T14:15:12 mega1: preferably in a pm 2010-09-19T14:19:03 *** justin_pdx has joined #aichallenge 2010-09-19T14:19:11 who can I contact to check why my bot fails at turn 1 (only on the official site)? couldn't find any contact link anywhere. 2010-09-19T14:22:42 *** dhartmei has quit IRC (Quit: This computer has gone to sleep) 2010-09-19T14:24:20 *** justin_pdx has quit IRC (Quit: justin_pdx) 2010-09-19T14:25:45 *** pc has quit IRC (Remote host closed the connection) 2010-09-19T14:25:53 *** pc has joined #aichallenge 2010-09-19T14:26:59 *** user42 has quit IRC (Ping timeout: 245 seconds) 2010-09-19T14:28:55 *** Yoshi-TS4 has quit IRC (Ping timeout: 272 seconds) 2010-09-19T14:34:14 dhartmei: stalemate 2010-09-19T14:36:21 got you that time 2010-09-19T14:36:41 congrats ;) 2010-09-19T14:39:06 Utkarsh: what programming language? 2010-09-19T14:50:57 *** craigferguson has joined #aichallenge 2010-09-19T14:52:13 *** TapewormGalactos has joined #aichallenge 2010-09-19T14:52:59 *** lotek has joined #aichallenge 2010-09-19T14:53:08 *** craigferguson has quit IRC (Client Quit) 2010-09-19T14:53:19 hi there 2010-09-19T14:53:32 quick question about the maps: are they always symmetrical? 2010-09-19T14:53:48 and do they have certain dimensions? 2010-09-19T14:54:34 symmetrical, yes 2010-09-19T14:56:11 Top 10 players: kcm1700(3513), McLeopold(3466), paulus(3397), adumlah(3392), EBraun(3329), montanalow(3310), golfin(3308), pdbpdb80(3255), wagstaff(3233), Loks1982(3230) 2010-09-19T14:57:57 wow, i lose to copybot 2010-09-19T14:57:58 http://72.44.46.68/canvas?game_id=46258 2010-09-19T14:58:10 not sure how that was possible 2010-09-19T14:58:33 bartwe: lol, fail 2010-09-19T14:59:32 turn 10, you send 33 ships from the lower left planet to the lower right 2010-09-19T15:00:00 slightly earlier you sent 37 from ll to ul 2010-09-19T15:00:19 so copybot did the same, and you lost both ul and lr 2010-09-19T15:01:10 i think that's copybot's fastest non-(D/I/T) win 2010-09-19T15:01:37 i seem to have fixed the invalid order problem 2010-09-19T15:01:44 so now copybot just loses a lot 2010-09-19T15:03:46 *** kozlovsky has joined #aichallenge 2010-09-19T15:14:22 teh suck 2010-09-19T15:17:22 hmmm this bot doesn't seem to suck as much once i've done a bit of tweaking 2010-09-19T15:17:33 i just uploaded it to official, so must be bed time 2010-09-19T15:17:34 night 2010-09-19T15:19:25 * Naktibalda made a site much nicer http://visualizer.naktibalda.lt/ 2010-09-19T15:25:16 Naktibalda: sorry, replying late. Ruby. 2010-09-19T15:27:24 *** mega1 has quit IRC (Ping timeout: 245 seconds) 2010-09-19T15:28:27 *** phreeza has joined #aichallenge 2010-09-19T15:35:27 pfft 2010-09-19T15:36:34 tiny bugs, maximizes the loss of growth rate by overcommiting on attack and trying to lure in counter attacks that are of negative value 2010-09-19T15:37:38 wait, how did "go" become an invalid instruction? 2010-09-19T15:38:27 copybot got a 0-turn loss due to invalid order 2010-09-19T15:40:24 zomg 2010-09-19T15:40:27 i won two games 2010-09-19T15:40:55 aparently maximizing loss is a bad stragtegy ;p 2010-09-19T15:41:32 lol 2010-09-19T15:42:02 wow the tcp server is way faster now 2010-09-19T15:42:40 Zannick: this means it sent anything after it sent 'go', but before got a 'go' from the server. 2010-09-19T15:42:44 *** Accoun has quit IRC () 2010-09-19T15:43:03 like, two 'go' immediately one after the other 2010-09-19T15:43:10 dhartmei_: weird 2010-09-19T15:43:31 my log showed the first turn data, followed by the server's "go", followed by its go 2010-09-19T15:43:47 *** misato has joined #aichallenge 2010-09-19T15:43:50 hi 2010-09-19T15:44:19 and then it sent another go: 2010.09.19 15:27:32 handle_line: player copybot: already moving: 'go' 2010-09-19T15:44:43 huh 2010-09-19T15:45:21 strange wouldn't expect somethign as simple as a copybot to have such an issue 2010-09-19T15:45:28 maybe the connection spazzed and tcp sent the message twice? 2010-09-19T15:45:37 Zannick: near impossible 2010-09-19T15:45:48 so is sending "go" twice 2010-09-19T15:46:10 copybot is nothing more than an implementation of "DoTurn" from the cpp starter code 2010-09-19T15:46:11 maybe a tcp capture with wireshark? 2010-09-19T15:46:30 serverside bug is a possibility too ofcourse 2010-09-19T15:46:55 it would be an improvement if the server would send an INFO or somesuch with the error message 2010-09-19T15:47:09 hmm, 17 different bots produced that error recently 2010-09-19T15:47:27 it does, the client is terminating before reading it, i have to fix that in tcp.c 2010-09-19T15:50:59 *** naglijo has joined #aichallenge 2010-09-19T15:53:48 *** RainCT has quit IRC (Quit: cya tomorrow!) 2010-09-19T15:54:39 *** Accoun has joined #aichallenge 2010-09-19T15:55:41 *** Naith has joined #aichallenge 2010-09-19T15:55:56 fixed a bug, uninitialized variable. restarting server. 2010-09-19T15:56:05 *** smellyhippy has quit IRC (Ping timeout: 252 seconds) 2010-09-19T15:56:07 it happened only on the first move. 2010-09-19T15:56:12 Top 10 players: McLeopold(3493), kcm1700(3488), paulus(3408), EBraun(3401), adumlah(3397), montanalow(3326), golfin(3318), wagstaff(3250), Loks1982(3234), pdbpdb80(3228) 2010-09-19T15:58:04 *** Bobng has joined #aichallenge 2010-09-19T16:00:51 *** smellyhippy has joined #aichallenge 2010-09-19T16:00:52 *** smellyhippy has joined #aichallenge 2010-09-19T16:02:58 *** Yoshi-TS4 has joined #aichallenge 2010-09-19T16:06:23 *** fisxoj has joined #aichallenge 2010-09-19T16:08:24 anything new? 2010-09-19T16:08:41 contestbot: seen accoun 2010-09-19T16:08:41 amstan: accoun was last seen in #aichallenge 2 days, 2 hours, 40 minutes, and 42 seconds ago: time to select starter pacage 2010-09-19T16:10:36 Is there any way to get a log from the contest server? My bot apparently crashed but I don't know if it's because of timeout or an actual crash. 2010-09-19T16:11:10 is pjreddie here? 2010-09-19T16:12:01 *** Fumpful has quit IRC (Ping timeout: 252 seconds) 2010-09-19T16:12:48 *** Appleman1234 has joined #aichallenge 2010-09-19T16:14:18 *** HakanD has joined #aichallenge 2010-09-19T16:14:55 *** justin_pdx has joined #aichallenge 2010-09-19T16:15:57 another question about the game mechanics: to you get new ships at the end or at the beginning of a turn? 2010-09-19T16:16:02 *do 2010-09-19T16:17:02 *** danielvf has quit IRC (Quit: danielvf) 2010-09-19T16:17:10 beginning 2010-09-19T16:17:59 so you do not get any income in the first turn then? 2010-09-19T16:19:37 *** ratatata has joined #aichallenge 2010-09-19T16:21:05 *** misato has quit IRC (Quit: Saliendo) 2010-09-19T16:21:14 and one other thing: what happens if two fleets (opposing forces) of same size attack the same neutral planet at the same time? 2010-09-19T16:21:31 *** ratatata has quit IRC (Client Quit) 2010-09-19T16:23:22 *** justin_pdx has quit IRC (Quit: justin_pdx) 2010-09-19T16:23:26 there is a detailed explanation about that on forums 2010-09-19T16:25:41 *** JCS^ has joined #aichallenge 2010-09-19T16:26:22 so the fleets of p1 and p2 just disappear and the planet stays neutral? 2010-09-19T16:26:24 TapewormGalactos: i think planet stays neutral 2010-09-19T16:26:26 yeah 2010-09-19T16:26:30 with 0 ships, but still neutral 2010-09-19T16:27:04 ok thx 2010-09-19T16:28:18 *** danielvf has joined #aichallenge 2010-09-19T16:33:35 *** Blkt has joined #aichallenge 2010-09-19T16:33:39 are there still maps with growthrate 0 planets on the server? 2010-09-19T16:38:44 *** tapwater has joined #aichallenge 2010-09-19T16:40:42 *** smf68 has joined #aichallenge 2010-09-19T16:44:47 *** customKoncept has joined #aichallenge 2010-09-19T16:47:04 any contest admins here (and awake) ? 2010-09-19T16:49:09 *** Bobng has quit IRC (Ping timeout: 240 seconds) 2010-09-19T16:50:26 *** jmreardon has quit IRC (Read error: Connection reset by peer) 2010-09-19T16:51:16 *** jmreardon has joined #aichallenge 2010-09-19T16:52:56 *** Queue29 has quit IRC (Remote host closed the connection) 2010-09-19T16:53:00 customKoncept: maybe, why? 2010-09-19T16:53:20 lost password 2010-09-19T16:53:32 sigh 2010-09-19T16:53:38 site or forum? 2010-09-19T16:53:43 site 2010-09-19T16:53:49 custom.koncept@gmail.com 2010-09-19T16:53:59 username? 2010-09-19T16:54:36 err.... username wasn't in the first 'validate your account' email 2010-09-19T16:54:48 and I think the other got swept into spam 2010-09-19T16:54:58 (always fun) 2010-09-19T16:55:15 so really, its username & password 2010-09-19T16:56:13 Top 10 players: McLeopold(3490), kcm1700(3464), EBraun(3441), paulus(3417), adumlah(3369), golfin(3312), montanalow(3311), wagstaff(3236), Loks1982(3229), dmj111(3220) 2010-09-19T16:57:33 customKoncept: the md5 of your password begins with be 2010-09-19T16:57:43 do you still want it changed or can you remember it? 2010-09-19T16:57:46 ahh 2010-09-19T16:57:57 changed, please 2010-09-19T16:58:02 customKoncept: the username is that same as here 2010-09-19T16:58:26 kk 2010-09-19T16:58:29 2 seconds 2010-09-19T16:58:38 might have the right password 2010-09-19T16:58:55 customKoncept: ok.. check your mail and reply to it, if you want it changed 2010-09-19T16:59:08 its fine 2010-09-19T16:59:09 thanks 2010-09-19T16:59:12 amstan, while you are at it, can you change my username from site, or delete it (: or any idea when this will be implemented? 2010-09-19T16:59:12 :) 2010-09-19T16:59:14 *** MetroStyles has joined #aichallenge 2010-09-19T16:59:41 HakanD: from what to what? and what's the email for it 2010-09-19T17:00:52 from HakanD -> Battal , email is hderyal [] gmail.com 2010-09-19T17:01:40 HakanD: first hex digit of the md5 hash of your password? 2010-09-19T17:02:06 wow http://72.44.46.68/canvas?game_id=48136 2010-09-19T17:02:10 sec 2010-09-19T17:02:34 *** hellman has quit IRC (Remote host closed the connection) 2010-09-19T17:07:40 *** Appleman1234 has quit IRC (Read error: Connection reset by peer) 2010-09-19T17:15:54 *** Appleman1234 has joined #aichallenge 2010-09-19T17:16:37 *** bourbaki has quit IRC (Remote host closed the connection) 2010-09-19T17:19:34 *** Eiii has joined #aichallenge 2010-09-19T17:21:12 *** Blkt has quit IRC (Ping timeout: 272 seconds) 2010-09-19T17:26:31 *** customKoncept has quit IRC (Quit: Page closed) 2010-09-19T17:31:23 *** nood has quit IRC (Quit: Leaving) 2010-09-19T17:33:57 *** sugarshark has quit IRC (Quit: ERC Version 5.3 (IRC client for Emacs)) 2010-09-19T17:41:57 *** TapewormGalactos has quit IRC (Ping timeout: 240 seconds) 2010-09-19T17:45:54 *** FlagCapper has quit IRC (Quit: Page closed) 2010-09-19T17:46:28 What the..! My bot keeps losing, even though it's clearly leading. http://ai-contest.com/visualizer.php?game_id=4588977 2010-09-19T17:46:38 Are there any logs? Output from the games? 2010-09-19T17:48:21 the minute your enemys planet has 0 ships on it you froze 2010-09-19T17:50:48 Cyndre: Hmm. You're right. That could definitely be a lead. 2010-09-19T17:51:51 Cyndre: Thanks, I'll look in to that. But it doesn't rhyme with my the one other game I lost: http://ai-contest.com/visualizer.php?game_id=4587312 2010-09-19T17:52:29 yea it does 2010-09-19T17:53:21 you have a neutral planet on the left side in the middle - blue red blue neutral (0 ships) 2010-09-19T17:54:07 somewhere your dividing by planetShips 2010-09-19T17:54:14 Cyndre: Nice catch. I missed that one. And you could just have made it into my list of heroes for the day. :) Thanks! 2010-09-19T17:54:22 thanks :) 2010-09-19T17:54:33 exception 'Exception' with message 'Num ships: 0.5' 2010-09-19T17:55:04 sending floats 2010-09-19T17:55:31 Sir we keep crashing. Dammit, its a war not a parade - send some fleets instead of floats 2010-09-19T17:56:00 anyways - thats some insight into my messed up mind :) 2010-09-19T17:56:13 Top 10 players: McLeopold(3495), kcm1700(3415), paulus(3402), EBraun(3380), adumlah(3358), golfin(3308), montanalow(3305), Loks1982(3273), wagstaff(3239), dmj111(3197) 2010-09-19T17:56:16 Naktibalda: You're seeing that message from my bot somewhere? 2010-09-19T17:56:36 Send fleets, not floats. Haha. 2010-09-19T17:57:29 *** justin_pdx has joined #aichallenge 2010-09-19T17:58:45 Yup. Think I just found the division with zero problem. 2010-09-19T17:59:40 I love patterns 2010-09-19T18:00:23 Would be nice to see the output from the game though, so one could spot these problems more easily. 2010-09-19T18:00:36 thats what the visualizer is :) 2010-09-19T18:01:55 Well, I'm just thinking my Python code probably crashed with a stacktrace, which would have been immensely helpful to me. 2010-09-19T18:03:42 yup, but gotta use the tools we want or patch the code and propoe a merge 2010-09-19T18:03:49 er tools we have 2010-09-19T18:03:54 propose a merge 2010-09-19T18:04:09 it is open source so you could get that into the server where it mails you a message on crash 2010-09-19T18:05:33 *** xvinyl has joined #aichallenge 2010-09-19T18:05:56 Don't know Java though. :/ 2010-09-19T18:06:16 *** pc has quit IRC (Remote host closed the connection) 2010-09-19T18:06:26 But yeah, I know. Less whining and more doing. :) 2010-09-19T18:08:58 *** justin_pdx has quit IRC (Quit: justin_pdx) 2010-09-19T18:09:51 *** coventry has joined #aichallenge 2010-09-19T18:12:01 or work with tools you have :) 2010-09-19T18:12:06 It'd be kind of cool to wrap the bot in something which used the information content of the game itself to transmit stack trace info or the like on the bot's failure. If you don't care about getting your ass kicked, you could communicate a lot of information through the fleet assignments... 2010-09-19T18:12:42 *** justin_pdx has joined #aichallenge 2010-09-19T18:15:14 I really need to finish rewritting my bot 2010-09-19T18:16:54 I just rewrote the awful python sample bot thing 2010-09-19T18:18:06 coventry: what information could you get that way that you could not get from simulating the game locally by feeding your bot a replay of its states? 2010-09-19T18:18:36 interesting http://ai-contest.com/forum/viewtopic.php?f=17&t=553&start=20#p4290 2010-09-19T18:18:51 also: the fleet assignments you can make without instantly losing change based on game state. designing a way to transmit information through them would not be easy 2010-09-19T18:18:54 *** pacotabaco has joined #aichallenge 2010-09-19T18:19:17 naglijo: not all bots are deterministic 2010-09-19T18:19:33 so simulating by feeding the states to your bot isn't always going to work 2010-09-19T18:20:28 at least not if you are hoping to get, say, information about server performance or something 2010-09-19T18:20:42 actually that applies even with a deterministic bot 2010-09-19T18:21:27 i think trying to get info about sevrer performance in that way would be violating the spirit of the rules. i cant imagine google would allow it. i assumed that's not what he meant 2010-09-19T18:21:56 i don't see what would in violation of the spirit of the rules about that 2010-09-19T18:22:17 the performance of the server drastically effects how you would write a search algorithm, for example 2010-09-19T18:22:41 naglijo: google doesn't run this contest, btw 2010-09-19T18:23:06 they just sponsor it in some vague, meaningless way 2010-09-19T18:23:49 would checking the time the enemy takes to make a move be considered cheating? 2010-09-19T18:23:59 :( my new version win 44/100 games against my old version 2010-09-19T18:24:32 Cyndre: how can you check? :) 2010-09-19T18:24:33 naglijo: Actually, *most* bots making full use of the resources will not be deterministic. The wall-time limit is random. 2010-09-19T18:24:55 Naktibalda: just checking time between moves and when you finished and when your new turn started 2010-09-19T18:25:00 s/random/fluctuating unpredictably/ 2010-09-19T18:25:22 you dont see what would violate the spirit of the rules by making moves that were not intended to win the game but rather to transmit information that would otherwise be impossible to get to you? 2010-09-19T18:26:02 "It's easier to beg forgiveness than ask permission." :-) 2010-09-19T18:26:14 the point being is if you can see they are taking longer maybe try a flood attack where you unleash 1 fleet to every planet from every 1 of your planets trying to make them go based on less information or even timeout 2010-09-19T18:26:18 Cyndre: both bots process simultaneously 2010-09-19T18:26:36 yes, i see what you mean about bots not being deterministic and so not being able to replay the game to get stack trace/state info from them 2010-09-19T18:26:37 so you would have to time between receiving game states, not between sending orders and receiving the next state 2010-09-19T18:27:14 naglijo: all games before the tournament are for testing anyway. testing the server seems perfectly fine to me 2010-09-19T18:28:20 jmcarthur: yes - if we start at 0 seconds, and I finish at .1 seconds and my next turn starts on .7 seconds - and then 1 turn dont say go until the time has been reached to time the inbetween time 2010-09-19T18:28:58 *** justin_pdx has quit IRC (Quit: justin_pdx) 2010-09-19T18:29:21 jmcarthur: and then you could flood fleets and time it to see if it takes more time etc 2010-09-19T18:29:34 i bet very few bots would actually be all that harmed by it 2010-09-19T18:29:37 *** phreeza has quit IRC (Quit: Computer has gone to sleep) 2010-09-19T18:30:08 *** javagamer has joined #aichallenge 2010-09-19T18:30:22 that's certainly not something that would hurt mine, and mine is already going to be very dependent on how much cpu time it actually gets 2010-09-19T18:30:49 you'll only get to "flood" with 23 fleets per turn anyway now that the server does fleet merging 2010-09-19T18:30:52 I've made a starter package which works great on my machine, but it doesn't compile for the contest. Could I ask an admin to check why it won't compile (since e-mails about that are down)? 2010-09-19T18:30:59 and that's if you own all planets already :P 2010-09-19T18:31:08 jmcarthur: no, send 1 fleet to each planet from each one of your planets 2010-09-19T18:31:10 hm 2010-09-19T18:31:18 the map is allowed to have comments in it, isnt it? 2010-09-19T18:31:18 500 fleets 2010-09-19T18:31:18 Cyndre: oh i see. so 23*23 fleets 2010-09-19T18:31:24 still not that many 2010-09-19T18:31:24 are moves also allowed to have comments in them? 2010-09-19T18:31:50 they are 2010-09-19T18:31:58 naglijo: the spec left that intentionally unstated 2010-09-19T18:31:59 jmcarthur: then do it for 3 turns etc 2010-09-19T18:32:06 jmcarthur: or 7 turns 2010-09-19T18:32:23 can you see your bot's actual stdout stream from the prod server, or do they only show you the replay visualization script's output? 2010-09-19T18:32:24 Cyndre: the other bot should be doing its own fleet merging anyway, and that can be better merging than the server can do 2010-09-19T18:32:26 but based on the timing of course to see how they respond 2010-09-19T18:32:32 Cyndre: since you don't really care about the sources of the fleets 2010-09-19T18:32:43 Cyndre: so really it's just going to be like 23 fleets per turn at max 2010-09-19T18:32:55 also, have people figured out how to get your stderr stream from PlayGame.jar (bundled with the starter packs?) 2010-09-19T18:33:07 i noticed it has a bug where it unintentionally 'eats' your stderr stream.. 2010-09-19T18:33:18 its an easy work around though 2010-09-19T18:33:19 naglijo: doesn't the new one not do that now? 2010-09-19T18:33:27 i didnt know there was a new one, woops. 2010-09-19T18:33:30 I am sure only a small percentage actually change how the fleets are handled and most will problably do a for x in pw.Fleets() type of thing 2010-09-19T18:33:44 i'm not as sure as you, is all 2010-09-19T18:34:33 but like I said - using the timing to check for coding weakness's 2010-09-19T18:34:39 it's a tactic that will probably work against a lot of badly coded bots, but will be actively harmful against better bots, so at best you would end up being the "best bot that still sucks" :P 2010-09-19T18:34:57 the timing trick may work if you aren't doing search 2010-09-19T18:34:58 Cyndre: when your opponent said 'go' isnt accessible to you 2010-09-19T18:34:59 *** coventry has left #aichallenge ("ERC Version 5.3 (IRC client for Emacs)") 2010-09-19T18:35:13 it would not work for me since i'm using up the entire time limit already 2010-09-19T18:35:21 so i can't time my opponent regardless 2010-09-19T18:35:28 jmcarthur: yes it would - I can gauge your decision to 2010-09-19T18:35:37 huh? 2010-09-19T18:35:39 jmcarthur: and see its depriving you of information 2010-09-19T18:35:53 i don't understand what information is being deprived... 2010-09-19T18:35:56 pw.Fleets() is nearly useless 2010-09-19T18:36:00 I agree 2010-09-19T18:36:40 Cyndre: i meant my bot couldn't *use* that tactic 2010-09-19T18:36:52 since it's using the time limit already 2010-09-19T18:36:58 oh 2010-09-19T18:37:00 jmcarthur: what's your username in contest? 2010-09-19T18:37:12 Naktibalda: jmcarthur, but i have no bot submitted yet 2010-09-19T18:37:49 Naktibalda: old ghc won't run it. i'm hoping we get new sandboxes so i can upload binaries or we at least get new servers with newer versions of software so i can at least use ghc 6.12.3... 2010-09-19T18:38:24 Naktibalda: it also isn't quite working yet ^_^ 2010-09-19T18:38:24 jmcarthur: what are the chances of that? 2010-09-19T18:38:43 delt0r_: the chances of new servers is high. the chances of a better sandbox is not as good 2010-09-19T18:38:59 1 out of 2... could be worse 2010-09-19T18:39:18 even a better sandbox is still feasible though 2010-09-19T18:39:29 if we can get it in KVM or something maybe 2010-09-19T18:39:40 how is this sandbox for java? is it still a bit performance hit? 2010-09-19T18:39:48 no performance hit currently 2010-09-19T18:39:55 because the sandbox really doesn't do anything all that great 2010-09-19T18:40:08 it's basically just chroot, resource limits, iptables rules, etc. 2010-09-19T18:40:10 no performance hit is good 2010-09-19T18:40:19 right... 2010-09-19T18:40:28 not really a "sandbox" 2010-09-19T18:40:31 yeah 2010-09-19T18:40:40 which is why everybody is iffy about arbitrary binaries 2010-09-19T18:40:54 I see the point of source submission 2010-09-19T18:40:56 although, frankly, compiling and running arbitrary source code isn't much better 2010-09-19T18:41:07 *** kozlovsky has quit IRC (Quit: Page closed) 2010-09-19T18:41:10 especially considering one could just use inline assembler or something 2010-09-19T18:41:20 Companys that i work with --money transfer app. was all source submission only... 2010-09-19T18:41:25 or any obfuscator 2010-09-19T18:41:27 ahh... 2010-09-19T18:41:38 but you *can* look at the code... 2010-09-19T18:41:38 requiring the source code really doesn't do anything 2010-09-19T18:41:44 if you wanted... 2010-09-19T18:41:47 sure you can, but that means nothing 2010-09-19T18:42:04 you could also use a disassembler and maybe even get better results than obfuscated source code 2010-09-19T18:42:06 sure... but the treat is aways there 2010-09-19T18:42:15 true 2010-09-19T18:42:31 you could use a dissembler on shit we compile too 2010-09-19T18:42:36 in java at least disassmbler can look better than the orginal source 2010-09-19T18:42:44 in that sense, the *interpreted* languages stand a better chance at obfuscation! 2010-09-19T18:42:57 delt0r_: yeah jvm bytecode is easy to decompile 2010-09-19T18:43:22 *** Appleman1234 has quit IRC (Quit: Leaving) 2010-09-19T18:43:57 dstufft: honestly, i know nobody would go for it, but i wish we would just allow binaries, even for this stupid not-quite-sandbox we have. no reason not to and it would mean i can code my bot however i want 2010-09-19T18:44:34 with a virtual servers it seems very doable 2010-09-19T18:44:48 yeah. hopefully we get that 2010-09-19T18:44:48 sandbox should be vm's with saved states and rollbacks 2010-09-19T18:44:53 dstufft: agreed 2010-09-19T18:45:48 there's a lot of the backend stuff which is really poorly done 2010-09-19T18:45:53 dstufft: i think qemu even allows you to specify the cpu cycles per virtual time elapsed, so the apparent performance of the VMs could even be made the same on differently speced machines 2010-09-19T18:46:16 you can tell it as a system evolved, it wasn't designed 2010-09-19T18:46:25 yeah 2010-09-19T18:46:33 a rewrite is probably a good idea for next time 2010-09-19T18:46:42 i got bored, and have a partial one lolz 2010-09-19T18:46:46 that is way generic 2010-09-19T18:46:50 yeah how's that going? 2010-09-19T18:47:00 do you have sandboxes? can they port over to this? 2010-09-19T18:47:02 :) 2010-09-19T18:47:26 Don't have the backend done yet, so far been mostly working with the frontend shit 2010-09-19T18:47:51 when i feel motivated to do it anyways 2010-09-19T18:48:04 *** hornairs has joined #aichallenge 2010-09-19T18:53:31 Which two example maps were nasty for playing against RageBot? 20 and..? 2010-09-19T18:53:54 *** agoandanon has quit IRC (Quit: Page closed) 2010-09-19T18:55:11 39? 2010-09-19T18:56:03 30? 2010-09-19T18:56:04 err no, 26 and 67 2010-09-19T18:56:13 Top 10 players: McLeopold(3504), kcm1700(3449), paulus(3400), EBraun(3382), adumlah(3370), montanalow(3343), golfin(3316), wagstaff(3285), Loks1982(3277), sighbot(3220) 2010-09-19T18:57:38 *** sigh has joined #aichallenge 2010-09-19T18:58:56 *** jorge__ has joined #aichallenge 2010-09-19T18:58:59 hi! 2010-09-19T18:59:06 hi 2010-09-19T19:02:08 if my bot compiles ok, bur Google gives an error, how can I debug that? (Java version) 2010-09-19T19:02:22 google? 2010-09-19T19:02:25 I've tried with java 1.6 and 1.5, my code compiles ok 2010-09-19T19:02:40 *** tobiassjosten has quit IRC (Quit: Leaving) 2010-09-19T19:02:40 I can't upload my code, it can't compile up there 2010-09-19T19:02:45 what do you mean by google 2010-09-19T19:02:56 contest site probably :) 2010-09-19T19:02:56 ah, google == ai challenge 2010-09-19T19:03:06 oh 2010-09-19T19:03:10 Google says "error while compiling submission" 2010-09-19T19:03:12 :/ 2010-09-19T19:03:23 that's it :) 2010-09-19T19:03:26 don't call it google 2010-09-19T19:03:34 what language is it? 2010-09-19T19:03:38 anyone working on getting lisp to compile properly? 2010-09-19T19:03:45 aha ok, the aicontest then 2010-09-19T19:03:50 google = google.com; ai contest = ai-contest.com ;) 2010-09-19T19:04:00 ok sorry :) 2010-09-19T19:04:04 heh, np 2010-09-19T19:04:05 Java 2010-09-19T19:04:32 I'm sure I'm going to get top100 with my improvements, so I'm very impatient to test it ;) 2010-09-19T19:04:35 jorge__: Did you try also running your bot using the tool? 2010-09-19T19:04:48 did you receive an email about compilation? 2010-09-19T19:04:50 yes, I can see battles here locally 2010-09-19T19:05:20 I received emails with correct compilations, but not new ones with the failure ones 2010-09-19T19:05:25 I didn't when it failed, but I got an email when upload was successful. funny, isn't it? 2010-09-19T19:05:41 the same than me 2010-09-19T19:06:01 so I have no clue what can happen to my code that the aicontest doesn't like :( 2010-09-19T19:07:13 Is it ok to change PlanetWars.java and Planet.java, etc., right? 2010-09-19T19:07:19 yea 2010-09-19T19:07:34 but MyBot.java must be the entry point 2010-09-19T19:07:45 is there any weird requisite? 2010-09-19T19:07:50 ok, it is 2010-09-19T19:07:52 MyBot is the static main thats invoked 2010-09-19T19:07:56 permissions? 2010-09-19T19:08:09 yes, I have it 2010-09-19T19:08:10 only if you write to disk... or use the network 2010-09-19T19:08:22 yes, I write to HD (logging) 2010-09-19T19:08:26 is that disallowed? 2010-09-19T19:08:29 you can't do that... however that wouldn't make it fail in compliation 2010-09-19T19:08:37 aha ok 2010-09-19T19:08:40 You can't do that 2010-09-19T19:08:46 I'll try without that 2010-09-19T19:08:51 No writing to HD --its in the rules 2010-09-19T19:09:15 What's the time limit on compiling? 2010-09-19T19:09:24 and... for logging, locally is there any other solution to log? 2010-09-19T19:09:30 jorge__: no 2010-09-19T19:09:40 :S 2010-09-19T19:09:43 You have to use TCP server or something 2010-09-19T19:09:53 ok 2010-09-19T19:09:55 its just too much for the server to do really 2010-09-19T19:10:12 I can dump several megs a game.... x100000 2010-09-19T19:10:15 thats a lot 2010-09-19T19:10:58 I think sockets are not allowed 2010-09-19T19:11:15 Could an admin please take a look at http://ai-contest.com/forum/viewtopic.php?f=18&t=732 ? 2010-09-19T19:11:24 Oh, you meant instead of the ai challenge server, nvm ;) 2010-09-19T19:11:56 They are using an old haskell version 2010-09-19T19:12:00 Info was on the forums 2010-09-19T19:12:20 Maybe test with that version? 2010-09-19T19:12:56 *** hornairs has quit IRC (Quit: hornairs) 2010-09-19T19:13:44 jorge__: Yea just to be clear. when i say TCP server i mean the TCP server ladder. where you run your bot locally .. You can't use tcp on the main site 2010-09-19T19:13:49 Ahh, OK. Let me take a look for that. 2010-09-19T19:14:06 Anyone know offhand what haskell version they are running? 2010-09-19T19:14:15 *** mog_ has quit IRC (Ping timeout: 252 seconds) 2010-09-19T19:14:16 delt0r_: ok thanks, I know about the tcp server (but I haven't used) 2010-09-19T19:17:14 * Naktibalda uploaded a version that logs to file, is it a bad thing? :) 2010-09-19T19:17:33 Naktibalda: It should barf 2010-09-19T19:17:56 it's working for 2 days 2010-09-19T19:18:03 did it barf? --oh 2010-09-19T19:18:25 are you the reason for the slow server >| 2010-09-19T19:18:41 ups 2010-09-19T19:21:34 *** iFire has quit IRC (Read error: Connection reset by peer) 2010-09-19T19:22:23 *** iFire has joined #aichallenge 2010-09-19T19:27:16 greghaynes: 6.8.2 2010-09-19T19:27:40 Olathe: ^^ 2010-09-19T19:29:49 is there some scheme (Racket) support on the server? 2010-09-19T19:30:00 Say a planet has a growth rate of 5 ships per turn. It has been reduced to 1 ship. On the next turn, a fleet of 5 enemy ships are going to hit it. Will the growth take effect before or after the fleet hits? 2010-09-19T19:30:12 Before. 2010-09-19T19:31:53 perfect 2010-09-19T19:33:11 tcp server is way faster btw 2010-09-19T19:33:20 jmcarthur: Ahh, thanks. 2010-09-19T19:33:46 *** justin_pdx has joined #aichallenge 2010-09-19T19:36:43 *** Eiii has quit IRC (Ping timeout: 240 seconds) 2010-09-19T19:38:35 *** xCalamity has joined #aichallenge 2010-09-19T19:38:36 *** choas has quit IRC (Quit: leaving) 2010-09-19T19:39:38 I'm trying a Java client 2010-09-19T19:39:51 a Java TCP client I saw on the forum 2010-09-19T19:40:01 but it crashes when the battle starts :( 2010-09-19T19:40:42 java TCP "72.44.46.68" "995" "javaexample" "java MyBot" "10000" 2010-09-19T19:41:03 what should be "javaexample" there? 2010-09-19T19:41:23 username 2010-09-19T19:41:36 ah, ok 2010-09-19T19:41:50 *** justin_pdx has quit IRC (Ping timeout: 240 seconds) 2010-09-19T19:41:52 just out of curiosity, how many lines of code has your bot? Mine is about 300 (python) 2010-09-19T19:42:08 INFO This is game_id=50511 2010-09-19T19:42:08 ERROR: unknown exception 2010-09-19T19:42:08 null 2010-09-19T19:42:14 *** delt0r___ has joined #aichallenge 2010-09-19T19:42:15 any idea? 2010-09-19T19:42:36 pacotabaco: mine is about 500 lines :) 2010-09-19T19:42:45 "java MyBot.jar" ? 2010-09-19T19:43:06 no, it's Mybot.class 2010-09-19T19:43:08 jorge__: If you'd like, try: http://ai-contest.com/forum/viewtopic.php?p=3836#p3836 2010-09-19T19:43:12 jorge__: tnks ;) 2010-09-19T19:43:12 I don't have the jar 2010-09-19T19:43:37 *** delt0r_ has quit IRC (Ping timeout: 255 seconds) 2010-09-19T19:46:59 >> Starting [java MyBot, 10000] 2010-09-19T19:46:59 [ERROR] Sorry, I could not run that bot program. 2010-09-19T19:47:01 i don't know a lot about java but... afak you need the jar 2010-09-19T19:47:07 ok, there mus be any error with my bot here 2010-09-19T19:47:16 but why can I run it locally then???? :( 2010-09-19T19:47:18 "java MyBot" is the command to run it if it's a class 2010-09-19T19:47:45 yes, it runs here 2010-09-19T19:47:51 I type "java MyBot" 2010-09-19T19:47:56 and it waits for the orders 2010-09-19T19:48:00 seems to be ok 2010-09-19T19:48:05 I write go 2010-09-19T19:48:08 and it answers go 2010-09-19T19:48:09 :) 2010-09-19T19:48:14 a previous version of tcp client don't accept parameters 2010-09-19T19:48:21 oh, you're trying to run tcp 2010-09-19T19:48:23 but, i think it's fixed now 2010-09-19T19:48:39 pacotabaco: what? 2010-09-19T19:48:43 yeah, make a script 2010-09-19T19:48:54 put "#!/usr/bin/sh" on the first line 2010-09-19T19:48:56 oh ok 2010-09-19T19:48:59 and "java MyBot" on the second 2010-09-19T19:49:58 yes i just check the code, it already has this "bug" 2010-09-19T19:50:24 yeah! it's playing! :D 2010-09-19T19:50:30 cool 2010-09-19T19:50:42 so why ai.contest doesn't accept my code then? 2010-09-19T19:50:54 ermm... 2010-09-19T19:50:56 Turn 34: 16 16 3 O 2010-09-19T19:50:56 >> Exiting because server closed connection 2010-09-19T19:51:06 the TCP client failed? or the server? 2010-09-19T19:52:31 what did you put as your username? 2010-09-19T19:53:02 spesgmd 2010-09-19T19:53:19 Starting [runMyBot.sh, 10000] 2010-09-19T19:53:20 >> Connecting to Planet Wars server at 72.44.46.68:995 as spesgmd 2010-09-19T19:53:20 >> You currently have -131 Elo 2010-09-19T19:53:36 you apparently issued an invalid order 2010-09-19T19:54:04 ahhh ok I see 2010-09-19T19:54:33 who's tcp server is that one? 2010-09-19T19:54:34 *** Utkarsh has quit IRC (Ping timeout: 276 seconds) 2010-09-19T19:54:46 (72.44.46.68:995) 2010-09-19T19:54:56 cool that TCP server BTW, thanks! :) 2010-09-19T19:55:24 sigh: same one, dhartmei_ has it on an ec2 instance now 2010-09-19T19:56:01 oh, so the old address is deprecated? 2010-09-19T19:56:14 Top 10 players: McLeopold(3516), kcm1700(3443), paulus(3394), EBraun(3377), adumlah(3352), montanalow(3339), golfin(3302), wagstaff(3274), Loks1982(3271), sighbot(3219) 2010-09-19T19:56:19 so far as I know 2010-09-19T19:56:22 thanks 2010-09-19T19:56:53 heh, scrapeing back in to the top 10 2010-09-19T19:59:08 grats 2010-09-19T19:59:15 still trying to crawl into double digits myself 2010-09-19T20:00:21 slowly keep making improvements and you'll get there :) 2010-09-19T20:00:32 *** Utkarsh has joined #aichallenge 2010-09-19T20:01:50 *** xCalamity has left #aichallenge ("Leaving") 2010-09-19T20:03:46 currently redoing everything from scratch 2010-09-19T20:04:49 ah yup, I did that on the weekend 2010-09-19T20:05:44 creating a game state I guess - list of every planet and the future turns upto 200 2010-09-19T20:06:15 I know I should probably ask a language specific channel, but ya'll are usually more helpful. In Java, if I have, say, List planets and List myPlanets (which was created by for(Planet p: planets){if(p.Owner() == 1) myPlanets.add(p)} , if I make a change to Planet in myPlanets, the same change is made in the corresponding planet in planets since they are the same object. Am I right? I apologize for the ignorance 2010-09-19T20:06:15 of this question, but I'm learning to program for this event. 2010-09-19T20:06:58 are you going to beat PHP bot? 2010-09-19T20:07:04 Naktibalda, I have been 2010-09-19T20:07:44 well, until friday anyway, then I stopped playing on the TCP server to work on ErrBot4. 2010-09-19T20:10:30 Cyndre: that's useful to have... I personally just calculate up the the arrival of the last fleet 2010-09-19T20:11:23 hmm, 200 turns is an idea 2010-09-19T20:12:24 yeah, a static 200 is simpler 2010-09-19T20:12:35 Cyndre, that's what I've been doing, but instead of going 200 turns out, I'm only going out the max distance number of turns and then each turn, I am creating a new turn at the end of my array 2010-09-19T20:13:21 * Naktibalda was thinking how to avoid recalculation at each turn, static array would be usefull for that proble, 2010-09-19T20:14:02 atm, I don't have issues with speed, so optimising away recalculation just isn't worth it for me 2010-09-19T20:14:42 I thought about doing the 200 turn array, but I thought it might get complicated since at each turn I would have to recalculate 200-n turns depending on what I or my enemy do. this way, I only have to calculate around 30 turns or so max 2010-09-19T20:15:00 at a time that is 2010-09-19T20:15:40 I hope the performance that I'll gain from that won't be overtaken by the cost of instantiating a new turn object every turn 2010-09-19T20:17:42 *** MetroStyles has quit IRC (Ping timeout: 252 seconds) 2010-09-19T20:19:00 * rogue780 is listening to crickets 2010-09-19T20:20:32 merely instantiating an object is pretty cheap in java iirc 2010-09-19T20:22:15 hopefully cheaper than modifying 200-n objects 2010-09-19T20:23:01 honestly, I don't see either being a performance bottleneck 2010-09-19T20:24:17 when you start running a lot of simulations then you need to start worrying about data-structures :) 2010-09-19T20:27:04 better data structures let simplify algorithms 2010-09-19T20:28:11 Naktibalda: simplifying algorithm and increasing performance are two different goals :P 2010-09-19T20:28:20 right 2010-09-19T20:28:32 ultimate goal is victory :) 2010-09-19T20:28:50 but yeah, currently I optimizing for flexibility, ease of refactoring and simplification of code 2010-09-19T20:28:59 *** Tnargav has joined #aichallenge 2010-09-19T20:32:28 *** Macuyiko has quit IRC (Ping timeout: 264 seconds) 2010-09-19T20:32:37 *** mightybyte has joined #aichallenge 2010-09-19T20:33:11 Hi there, how should a command in cmd window look like to try my bot in C++? There is only java example - which works for bots - but I can't run mine ;/ 2010-09-19T20:33:34 "bot.exe" 2010-09-19T20:34:06 so instead of "jaca mybot" type "bot.exe" right? 2010-09-19T20:34:11 java* 2010-09-19T20:34:14 or "./MyBot" if you're running linux 2010-09-19T20:34:17 Tnargav, yes 2010-09-19T20:37:56 Hm... I can't even compile the example from the website. Does anyone have good source code that compiles? 2010-09-19T20:41:12 starter pack should be good 2010-09-19T20:42:35 if it doesn't compile, you need another compiler 2010-09-19T20:42:35 >> Exiting because server closed connection 2010-09-19T20:42:45 when only a few turns, with the Java TCP client 2010-09-19T20:42:53 is that my fault? 2010-09-19T20:42:54 from what I see one file wasn't included. It compiled but only with few wornings now. Let's see if it works 2010-09-19T20:42:58 or server's? 2010-09-19T20:43:16 jorge__: what status do you see in a game list? 2010-09-19T20:43:59 jorge__: you have lost 3 games because of Invalid input 2010-09-19T20:44:02 invalid input 2010-09-19T20:44:04 :( 2010-09-19T20:44:36 how can I see the bad input? is that possible? 2010-09-19T20:44:44 log your orders 2010-09-19T20:44:50 yup 2010-09-19T20:45:09 yes, I see it locally, ok 2010-09-19T20:45:13 thanks... 2010-09-19T20:46:53 Ok it worked :) Thanks for a hint guys ^^ 2010-09-19T20:47:23 *** Theobon has quit IRC (Ping timeout: 240 seconds) 2010-09-19T20:54:29 *** Tnargav has quit IRC (Quit: Page closed) 2010-09-19T20:56:15 Top 10 players: McLeopold(3505), kcm1700(3450), paulus(3415), EBraun(3358), golfin(3310), adumlah(3306), wagstaff(3255), Loks1982(3253), dhartmei(3230), sighbot(3207) 2010-09-19T21:04:16 *** Yoshi-TS4 has quit IRC (Quit: Minus One IRC) 2010-09-19T21:12:22 *** Lithosphere has quit IRC (Remote host closed the connection) 2010-09-19T21:12:37 *** Lithosphere has joined #aichallenge 2010-09-19T21:16:32 position 109 - climbing towards double digits 2010-09-19T21:18:32 *** mightybyte has quit IRC (Quit: leaving) 2010-09-19T21:19:16 it's hard to find mistypes at 2am 2010-09-19T21:19:38 I had flights instead of fleets and I couldn't see it 2010-09-19T21:19:55 lol :P 2010-09-19T21:20:21 Cyndre, I'll hopefully be joining you later this week 2010-09-19T21:20:57 so hard trying to do it 2010-09-19T21:21:09 the biggest issue is going through all my old code - its why Im working on a new view 2010-09-19T21:21:25 but not sure if its worth while 2010-09-19T21:21:26 lol 2010-09-19T21:22:45 I completely started over on friday 2010-09-19T21:22:54 my previous bot was a dumbot 2010-09-19T21:23:06 it merely reacted to events in the current turn 2010-09-19T21:23:10 mine is pretty smart but the code is a mess and getting hard to work on 2010-09-19T21:23:14 *** sigh has quit IRC (Remote host closed the connection) 2010-09-19T21:23:24 Man, I'm fairly addicted to programming these stupid bots now! :( 2010-09-19T21:23:25 my new one will be a smartbot (i hope) 2010-09-19T21:23:33 *** sigh has joined #aichallenge 2010-09-19T21:24:09 Would anyone be interested in starting a site with me that holds different programming contests year-round? 2010-09-19T21:25:56 I'd be interested in playing all the contests on that site year round :P 2010-09-19T21:27:05 *** Lithosphere has quit IRC (Remote host closed the connection) 2010-09-19T21:27:22 *** Lithosphere has joined #aichallenge 2010-09-19T21:27:35 *** nibalizer has quit IRC (Ping timeout: 276 seconds) 2010-09-19T21:27:40 *** nibalizer has joined #aichallenge 2010-09-19T21:28:33 what about codecontest.com? 2010-09-19T21:28:39 decent domain? 2010-09-19T21:28:55 taken 2010-09-19T21:29:03 not taken 2010-09-19T21:29:09 unless you just took it 2010-09-19T21:30:14 hmm...register.com lied to me it seems 2010-09-19T21:30:40 nah I was just super quick with paypal 2010-09-19T21:30:42 oh I see. it's a premium domain available for $1888 2010-09-19T21:30:42 jks 2010-09-19T21:31:09 anyone have domain ideas? 2010-09-19T21:31:19 .net and .ord are free 2010-09-19T21:31:27 why not .org 2010-09-19T21:31:29 aicontest.org 2010-09-19T21:31:49 It's too close to ai-contest.com and I don't want to detract from it 2010-09-19T21:32:02 or be detracted from by it 2010-09-19T21:32:22 codeai.org? 2010-09-19T21:32:28 have you anything more in your plan than "lets have more contests"? 2010-09-19T21:32:43 http://codeai.com/ 2010-09-19T21:32:54 ^ LUSHES BROWS 2010-09-19T21:33:03 I knew computers would be useful one day 2010-09-19T21:33:19 luscious* 2010-09-19T21:33:51 lmao 2010-09-19T21:34:06 *** Cyndre has quit IRC (Quit: Leaving) 2010-09-19T21:34:22 brb. and yes i have more to the idea 2010-09-19T21:34:53 cracking the code :) 2010-09-19T21:34:56 *** jorge__ has quit IRC (Remote host closed the connection) 2010-09-19T21:36:36 what does "LUSHES BROWS" mean? 2010-09-19T21:36:58 *luscious brows ^^ 2010-09-19T21:38:23 what about challengeai.com? or does codeai.com sound better? 2010-09-19T21:39:09 I'm not sure if I want it to be strictly AI though. I'm envisioning differnet categories of challenges 2010-09-19T21:39:31 thefuckingbestcoder.com 2010-09-19T21:39:59 an AI one, of course, but also smallest program to do X challenge, or most effecient to do X challenge 2010-09-19T21:40:37 don't forget the most obfuscated code challenge 2010-09-19T21:41:15 unless it's a solved problem = perl allways wins 2010-09-19T21:41:33 brb. gotta take the kids to the pool 2010-09-19T21:41:37 rogue780: codeolympics.com 2010-09-19T21:42:06 Naktibalda: J will usually get you both most obfuscated and shortest 2010-09-19T21:42:06 holy shit. that's it 2010-09-19T21:42:28 i am the master of domains 2010-09-19T21:42:30 gold medal in coding on the run 2010-09-19T21:42:50 crap. where's my wallet 2010-09-19T21:43:19 * rogue780 is using his unemployment debit card for this most important purchase 2010-09-19T21:43:26 lol 2010-09-19T21:44:38 registered. now i am seriously taking the browns to the super bowl 2010-09-19T21:49:05 *** Cyndre has joined #aichallenge 2010-09-19T21:49:49 is there a windows tcp client yet? 2010-09-19T21:50:54 the tcp client doesn't work on windows? 2010-09-19T21:51:05 http://ai-contest.com/forum/viewtopic.php?f=18&t=424&start=40#p3786 2010-09-19T21:54:22 sigh, you use Windows :P 2010-09-19T21:54:24 FAIL 2010-09-19T21:54:36 sigh, i have one compiled 2010-09-19T21:54:40 I'm not using windows 2010-09-19T21:54:55 sigh, www.errgaming.com/tcp.zip has all the binary files you need to just run the tcp client and go 2010-09-19T21:55:09 lol, Windows :P 2010-09-19T21:56:15 Top 10 players: McLeopold(3532), kcm1700(3447), paulus(3417), EBraun(3373), adumlah(3306), golfin(3305), wagstaff(3218), dhartmei(3212), Loks1982(3198), sighbot(3194) 2010-09-19T21:56:27 meh, I'm confused... I don't use windows so I was just wondering why the normal tcp client doesn't work for Cyndre 2010-09-19T21:56:42 *** Theobon has joined #aichallenge 2010-09-19T21:58:25 * sigh retreats back into his shell 2010-09-19T21:58:30 graphitemaster: be nice to people, you're not in #sauerbraten anymore 2010-09-19T21:59:20 2010 will be the year of Linux on the desktop! 2010-09-19T21:59:40 though, i do agree with him, lol linux ftw 2010-09-19T22:01:45 ... 2010-09-19T22:01:52 Linux FTW! 2010-09-19T22:02:06 I say 2012 will be the year everyone is uses linux 2010-09-19T22:03:06 *** bradtgmuray has quit IRC (Quit: leaving) 2010-09-19T22:03:40 err, i guess my last was to Cyndre not sigh. reading fail 2010-09-19T22:04:23 hmm, actually... we could have a set of monad instances for several sizes of tuples 2010-09-19T22:04:55 instance Monoid a => Monad ((,) a), instance (Monoid a, Monoid b) => Monad ((,,) a b), etc. 2010-09-19T22:05:50 * rogue780 did not enjoy sauerbraten 2010-09-19T22:06:08 *** Theobon has quit IRC (Ping timeout: 265 seconds) 2010-09-19T22:06:16 jmcarthur: you meant to post in #haskell? 2010-09-19T22:06:42 *** justin_pdx has joined #aichallenge 2010-09-19T22:06:51 rogue780: why not? 2010-09-19T22:07:01 sigh: dang, i did 2010-09-19T22:07:07 *** Zarathust has quit IRC (Quit: Page closed) 2010-09-19T22:07:32 *** Zarathust has joined #aichallenge 2010-09-19T22:07:49 amstan, IIRC, i felt it was too fast paced and impossible to play 2010-09-19T22:08:06 rogue780: takes a while to get used to 2010-09-19T22:08:09 rogue780: but it's fun 2010-09-19T22:11:15 my bot is becoming worse after each improvement :( 2010-09-19T22:13:35 those aren't called improvements 2010-09-19T22:13:59 :) 2010-09-19T22:14:06 Naktibalda, I went through that phase. 2010-09-19T22:14:55 *** Naktibalda has quit IRC (Remote host closed the connection) 2010-09-19T22:16:19 amstan, & rogue780 what game are you talking about? 2010-09-19T22:17:30 graphitemaster: sauer.. 2010-09-19T22:17:47 figures, that game is awesome 2010-09-19T22:18:10 contestbot: seen j3camero 2010-09-19T22:18:10 amstan: j3camero was last seen in #aichallenge 3 days, 23 hours, 27 minutes, and 59 seconds ago: Cyndre: are you interested in building it? That would be soooo bitchin. I have seen some simple online IDEs before, and I gotta say it's pretty cool. We could build something unprecedented! 2010-09-19T22:18:18 when is he ever coming here? 2010-09-19T22:18:26 *** mightybyte has joined #aichallenge 2010-09-19T22:19:21 so... does everything have to be in your base directory when you submit your bot? 2010-09-19T22:19:36 DirtyKeyboard: yes, as you found it in the starter package 2010-09-19T22:19:50 amstan: gah, that's annoying 2010-09-19T22:19:51 DirtyKeyboard: check the repo if you actually want to see the code that opens up your package 2010-09-19T22:20:07 DirtyKeyboard: folders might work, but your main file needs to be in root 2010-09-19T22:20:17 *** naglijo_ has joined #aichallenge 2010-09-19T22:20:28 DirtyKeyboard: you're welcome to provide a patch 2010-09-19T22:20:53 *** naglijo has quit IRC (Ping timeout: 276 seconds) 2010-09-19T22:25:48 *** Satya has joined #aichallenge 2010-09-19T22:29:41 for you guys talking about your bot not doing as well as you 'improve it' I'm under the same thing right now... it's a little frustrating 2010-09-19T22:30:35 I solved that problem by writing bugs 2010-09-19T22:30:41 often they make my program better 2010-09-19T22:31:00 but... I don't write bugs 2010-09-19T22:31:02 :D 2010-09-19T22:31:16 okay, that's a bold faced lie 2010-09-19T22:31:22 lol 2010-09-19T22:37:09 nothing wrong with windows - right tool for the right job 2010-09-19T22:38:33 and already usikng the tcp.c on my server downstairs, but its not as nice as my windows machine 2010-09-19T22:38:52 Cyndre, did you try my compiled binary? 2010-09-19T22:39:06 no, didnt see the link for it 2010-09-19T22:39:14 oh - www.errgaming.com/tcp.zip 2010-09-19T22:39:16 ? 2010-09-19T22:39:17 www.errgaming.com/tcp.zip 2010-09-19T22:39:19 yup 2010-09-19T22:39:27 windows 7? 2010-09-19T22:39:30 yes 2010-09-19T22:39:38 32bit 2010-09-19T22:39:52 compiled with cygwin 2010-09-19T22:40:05 anyway, im going to bed 2010-09-19T22:40:08 g'night 2010-09-19T22:40:13 *** rogue780 is now known as rogue780|away 2010-09-19T22:41:22 *** pc has joined #aichallenge 2010-09-19T22:41:25 *** xvinyl has quit IRC (Remote host closed the connection) 2010-09-19T22:44:10 *** jmreardon has quit IRC (Quit: jmreardon) 2010-09-19T22:45:21 *** justin_pdx has quit IRC (Quit: justin_pdx) 2010-09-19T22:45:24 *** Eiii has joined #aichallenge 2010-09-19T22:47:16 *** jmreardon has joined #aichallenge 2010-09-19T22:48:22 *** pc is now known as xvinyl 2010-09-19T22:48:38 *** jmreardon has quit IRC (Client Quit) 2010-09-19T22:49:58 *** wh1teside has quit IRC (Quit: bbiab) 2010-09-19T22:51:05 *** xvinyl has quit IRC (Remote host closed the connection) 2010-09-19T22:51:37 *** hornairs has joined #aichallenge 2010-09-19T22:55:02 *** DjinnKahn has joined #aichallenge 2010-09-19T22:56:15 Top 10 players: McLeopold(3526), kcm1700(3462), EBraun(3404), paulus(3378), golfin(3352), adumlah(3314), sighbot(3241), Titannet1(3226), wagstaff(3214), locutus(3203) 2010-09-19T22:58:40 *** pacotabaco has quit IRC (Quit: Page closed) 2010-09-19T23:03:09 my new bot - just uploaded 2010-09-19T23:03:10 http://ai-contest.com/visualizer.php?game_id=4594274 2010-09-19T23:04:56 you are red? 2010-09-19T23:05:35 tea 2010-09-19T23:05:37 er yea 2010-09-19T23:05:57 I need to fix my reinforcement code 2010-09-19T23:06:07 shouldnt be over reinforcing 2010-09-19T23:06:11 cool, now you just have to wait a million hours settle at your rank :P 2010-09-19T23:06:42 *** deepblue has joined #aichallenge 2010-09-19T23:06:58 * deepblue yawns 2010-09-19T23:07:07 yeah, seems like you keep a constant base of 20-30 on any planet 2010-09-19T23:07:15 that's a bit conservative 2010-09-19T23:07:16 good evening guys 2010-09-19T23:07:21 hi 2010-09-19T23:07:23 is it? 2010-09-19T23:08:12 depends on the position, but I've found a hard limit is generally not good 2010-09-19T23:09:54 thats part of the reinforcement code anyways 2010-09-19T23:10:12 i have a hard time getting my bod to not stationed too many stuff in the back when it's needed at the front 2010-09-19T23:10:33 my planets at the back always send out full growth every turn to closer planets 2010-09-19T23:10:59 I have done too complete rewrite since last weekend 2010-09-19T23:11:00 well, not always... they do defend themselved 2010-09-19T23:11:09 one was an utter failure :/ 2010-09-19T23:11:16 it happens 2010-09-19T23:11:28 yes yes 2010-09-19T23:11:33 all good and fun :) 2010-09-19T23:11:36 :) 2010-09-19T23:12:01 *** hornairs has quit IRC (Ping timeout: 245 seconds) 2010-09-19T23:14:48 sigh, how do you calculate whether or not a planet should send reinforcments or attack? I mean, how do you determine what planets are "in the back" vs the ones "in the front" and the ones in the middle?\ 2010-09-19T23:14:53 *** rogue780|away is now known as rogue780 2010-09-19T23:15:55 rogue780: attacks are based on my planet evaluation function... I don't want to say what that is right now 2010-09-19T23:16:20 (my policy on strategy advice is that I am happy to say anything taht I think is common knowledge to the top 10-20 players) 2010-09-19T23:16:33 rogue780: average of the distances to the enemy planets? 2010-09-19T23:16:35 come on...it's just between you and me, and I promise I won't use it... 2010-09-19T23:16:39 :) 2010-09-19T23:16:59 rogue780: scroll to the bottom, say cheese: http://contestbot.hypertriangle.com/channellogs/%23aichallenge/%23aichallenge.2010-09-19.log 2010-09-19T23:17:34 amstan: that sounds easy in theory, but it is difficult ot put in practice... at least for me 2010-09-19T23:17:43 sigh: why? 2010-09-19T23:18:02 sigh: shouldn't take that long 2010-09-19T23:18:24 what if the enemy is in separate clusters, how to you count planets that are owned by one player now, but another later? 2010-09-19T23:18:44 hmm...maybe if I divide the map into a 5x5 grid it could aid with analysis 2010-09-19T23:19:12 what if an enemy takes a planet right in the middle of an area you control? 2010-09-19T23:19:31 sigh, these are the exact problems I've been running into 2010-09-19T23:19:45 yeah, it's not a trivial problem for me 2010-09-19T23:19:51 gar! 2010-09-19T23:19:56 maybe amstan has an awesome solution :) 2010-09-19T23:20:04 i do? 2010-09-19T23:20:19 I've been trying to figure out a reinforcement strategy, but nothing has panned out. I'm hoping my new approach will make it easier 2010-09-19T23:21:06 rogue780: this is one thing where it definitely helps to watch the top games 2010-09-19T23:21:11 my problem is how to make it cautiously aggressive and not cautiously passive ;) 2010-09-19T23:21:32 my bot doesn't know the difference between defense and offense 2010-09-19T23:21:48 reinforcement strategies are slightly easier to figure out than other strategies when watching games 2010-09-19T23:21:51 exactly my problem 2010-09-19T23:22:03 mine is by design 2010-09-19T23:22:15 the line is kind of blurred for planets that are under attack 2010-09-19T23:22:21 so it's not good to make a clear distinction 2010-09-19T23:22:32 also you don't want ot fall for the sunk-cost fallacy 2010-09-19T23:24:16 yeah, my bot doesn't even care about who owns what planet, actually 2010-09-19T23:24:26 :) 2010-09-19T23:24:35 like, it's implicitly considered, but it's not an explicit concept 2010-09-19T23:25:27 mine does somehow 2010-09-19T23:25:36 it does like to kill the enemy 2010-09-19T23:25:39 at least not in the actual AI part of it. it's concrete in the core simulation <-- gives away my approach a bit 2010-09-19T23:25:50 hehe 2010-09-19T23:26:05 what is the name of your bot? 2010-09-19T23:26:10 it's not up yet 2010-09-19T23:26:15 :) 2010-09-19T23:26:16 hmmm... I don't have a simulation yet... only prediction of future states 2010-09-19T23:26:17 or on tcp 2010-09-19T23:26:30 not on tcp yet 2010-09-19T23:26:30 ? 2010-09-19T23:26:33 prediction is just a simulation heuristic ;) 2010-09-19T23:26:37 no not yet 2010-09-19T23:26:41 k 2010-09-19T23:27:00 it'll get there. i'm just taking my time 2010-09-19T23:27:07 i mean... prediction of states based on current fleet, not prediction of opponent's moves 2010-09-19T23:27:09 sure 2010-09-19T23:27:09 and my idea has many interdependent parts 2010-09-19T23:27:24 *** blbrown_win3 has joined #aichallenge 2010-09-19T23:27:27 I'm trying to modularize my approach 2010-09-19T23:27:43 do the start packages get updates. Do I ever need to update during the course of the contest 2010-09-19T23:27:47 for example, my redistribution code is independent of my attack code 2010-09-19T23:27:53 blbrown_win3: maybe 2010-09-19T23:27:55 my indivdual components can be tested in isolation, but they don't form a complete bot except when composed together 2010-09-19T23:28:02 *individual 2010-09-19T23:28:08 blbrown_win3: what language? 2010-09-19T23:28:15 amstan, java 2010-09-19T23:28:27 blbrown_win3: so for starter packages, once they're done their done 2010-09-19T23:28:31 the protocol stays fixed 2010-09-19T23:28:36 ok 2010-09-19T23:28:38 though, there might be bugs in them 2010-09-19T23:28:47 components? 2010-09-19T23:28:49 i am in fact very likely to write unit tests for some of this code :o 2010-09-19T23:28:52 mine has two functions :D 2010-09-19T23:29:08 I am writing a language for my bot ... for some strange reason 2010-09-19T23:29:11 my unit tests were endless time in gdb :/ 2010-09-19T23:29:15 yeah, like i said, i actually have some code to write :P 2010-09-19T23:29:19 ew 2010-09-19T23:29:27 I know 2010-09-19T23:29:31 i might 2010-09-19T23:29:31 mine will be properties tested by quickcheck! :D 2010-09-19T23:29:40 rip it all apart in testable chunks one day 2010-09-19T23:30:03 what are unit tests? 2010-09-19T23:30:05 quickcheck is awesome 2010-09-19T23:30:05 j/k :P 2010-09-19T23:30:21 my code is just what comes out when I spend half a night just writing stuff down without much testing 2010-09-19T23:30:33 horrible +/- 1 errors and stuff :D 2010-09-19T23:30:46 off by 1 errors can kill in this game 2010-09-19T23:30:48 i'd bring lambdabot in here to show off quickcheck, but contestbot would just get noisy since it looks for the @ syntax 2010-09-19T23:31:07 contestbot vs lambdabot: FIGHT 2010-09-19T23:31:08 sigh: An error occured while trying to show the previous error. 2010-09-19T23:31:14 :P 2010-09-19T23:31:25 yeah off by 1 is the difference between barely winning a battle and losing it completely 2010-09-19T23:31:31 lol at that error message 2010-09-19T23:31:46 sigh: with unit tests you for example check a function if it produces the correct output for a given input 2010-09-19T23:31:47 *** McLeopold has joined #aichallenge 2010-09-19T23:31:57 deepblue: i think sigh knows what a unit test is :P 2010-09-19T23:32:02 hehe 2010-09-19T23:32:05 ok ok 2010-09-19T23:32:06 :D 2010-09-19T23:32:11 I shall shut up then 2010-09-19T23:32:22 but a minor correction: 2010-09-19T23:32:26 I missed the irony tag 2010-09-19T23:32:31 jmcarthur: hold on.. 2010-09-19T23:32:38 a unit test doesn't necessarily have to check for a specific input 2010-09-19T23:32:40 contestbot: list config supybot 2010-09-19T23:32:40 amstan: (list [--private] []) -- Lists the commands available in the given plugin. If no plugin is given, lists the public plugins available. If --private is given, lists the private plugins. 2010-09-19T23:32:47 contestbot: config list supybot 2010-09-19T23:32:47 amstan: @abuse, @capabilities, @commands, @databases, @debug, @directories, @drivers, @log, @networks, @nick, @plugins, @protocols, @replies, @reply, alwaysJoinOnInvite, defaultIgnore, defaultSocketTimeout, externalIP, flush, followIdentificationThroughNickChanges, ident, pidFile, snarfThrottle, upkeepInterval, and user 2010-09-19T23:33:06 hmm 2010-09-19T23:33:12 it's okay 2010-09-19T23:33:18 *** pjreddie has joined #aichallenge 2010-09-19T23:33:19 contestbot: help botdoc 2010-09-19T23:33:19 amstan: Error: There is no command "botdoc". 2010-09-19T23:33:23 contestbot: botdoc 2010-09-19T23:33:24 amstan: botdoc = (#1) http://ubottu.com/stdin/supydocs/plugins/, or (#2) http://supybook.fealdia.org/devel/, or (#3) http://sourceforge.net/apps/mediawiki/gribble/index.php?title=Supybot_Resources 2010-09-19T23:33:42 *** mightybyte has quit IRC (Quit: mightybyte) 2010-09-19T23:33:42 contestbot: config reply.whenAddressedBy.chars 2010-09-19T23:33:42 amstan: Global: @; #aichallenge: @ 2010-09-19T23:33:45 contestbot: config reply.whenAddressedBy.chars ! 2010-09-19T23:33:45 amstan: Error: You don't have the owner capability. If you think that you should have this capability, be sure that you are identified before trying again. The 'whoami' command can tell you if you're identified. 2010-09-19T23:34:06 contestbot: config reply.whenAddressedBy.chars ! 2010-09-19T23:34:06 amstan: Ready to serve. 2010-09-19T23:34:09 there we go 2010-09-19T23:34:11 !yay 2010-09-19T23:34:11 Double rainbow all the way! 2010-09-19T23:34:13 jmcarthur: now you can bring it 2010-09-19T23:34:14 awesome 2010-09-19T23:34:37 *** lambdabot has joined #aichallenge 2010-09-19T23:34:40 what will lambdabot do? 2010-09-19T23:34:42 geeze, all that just to get the double rainbow? 2010-09-19T23:34:42 jmcarthur: but why? lol 2010-09-19T23:34:44 other than speak haskell 2010-09-19T23:35:16 a lot of things. i was only mentioning showing off quickcheck, but i wasn't actually intending to ask amstan to do that :P 2010-09-19T23:35:45 jmcarthur: you could have your own version here, just have some other command char 2010-09-19T23:35:52 *** hornairs has joined #aichallenge 2010-09-19T23:36:56 *** danielvf has joined #aichallenge 2010-09-19T23:37:09 @check \xs ys -> length (zip xs ys) == (length xs) `min` (length ys) 2010-09-19T23:37:11 "OK, passed 500 tests." 2010-09-19T23:37:18 @check \xs ys -> length (zip xs ys) == length xs 2010-09-19T23:37:20 "Falsifiable, after 3 tests:\n[(),()]\n[]\n" 2010-09-19T23:37:32 and that's all. totally worth the trouble 2010-09-19T23:37:38 lol 2010-09-19T23:37:40 *** danielvf has left #aichallenge 2010-09-19T23:37:55 jmcarthur: oh yeah.. i can make my bot have different chars for each channel 2010-09-19T23:37:56 *** lambdabot has left #aichallenge 2010-09-19T23:38:02 contestbot: config reply.whenAddressedBy.chars !@# 2010-09-19T23:38:02 amstan: I come to serve. 2010-09-19T23:38:19 !@#sigh 2010-09-19T23:38:20 sigh: An error occured while trying to show the previous error. 2010-09-19T23:38:20 i don't know how to do it with lambdabot or if that's even possible 2010-09-19T23:38:46 jmcarthur: most config entries from supybot can be specific to one channel only 2010-09-19T23:39:01 oh yeah.. anyone getting the referrence from the success messages? 2010-09-19T23:39:02 @success 2010-09-19T23:39:02 amstan: Ready to serve, my lord. 2010-09-19T23:39:11 lol 2010-09-19T23:39:13 will turns remaining reach 0 when the fleet hits the planet, or will the lowest number it reaches be 1, and if so would the be the turn that it hits or the turn before it hits? 2010-09-19T23:39:51 you won't be given fleets with turns remaining = 0 2010-09-19T23:39:52 On the server side, it should hit zero, do battle, then disappear 2010-09-19T23:40:06 yes 2010-09-19T23:40:17 so I'll never see it hit 0 then 2010-09-19T23:40:20 so if you spawn a fleet with distance 1 2010-09-19T23:40:25 exactly 2010-09-19T23:40:34 thanks. that is very important to know 2010-09-19T23:40:52 someone said that distance 1 planets weren't going to be allowed? 2010-09-19T23:41:08 send state, receive orders, depart, advance and grow, fight, check end conditions 2010-09-19T23:41:12 they should make a map with a chain of distance one planets :) 2010-09-19T23:41:30 those steps are repeated exactly n times, for n = number of turns 2010-09-19T23:41:40 stealth battles! 2010-09-19T23:41:52 and some of those steps can also check for end conditions on their own (errors) 2010-09-19T23:42:10 *** Appleman1234 has joined #aichallenge 2010-09-19T23:42:13 well, time to get to bed or my wife will kill me 2010-09-19T23:42:19 *** rogue780 is now known as rogue780|away 2010-09-19T23:42:29 hehe 2010-09-19T23:43:45 *** ermau has quit IRC (Quit: ermau) 2010-09-19T23:47:47 *** drewrobb has joined #aichallenge 2010-09-19T23:48:41 *** danielvf_ has joined #aichallenge 2010-09-19T23:53:09 *** FlagCapper has joined #aichallenge 2010-09-19T23:53:57 If I submitted RandomBot to the official server, would my upcoming losses drag down the rating of everyone I've beat? 2010-09-19T23:54:18 your rating is wiped when you upload a new bot 2010-09-19T23:54:28 it will work on tcp though :P 2010-09-19T23:54:51 Yeah, my rating. But what about other people's rating? 2010-09-19T23:55:05 *** DirtyKeyboard has quit IRC (Quit: DirtyKeyboard) 2010-09-19T23:55:07 Would they be considered to have lost to a 1500 bot? 2010-09-19T23:55:36 i assume the game of your old bot no longer count... but i'm just guessing 2010-09-19T23:55:41 *games 2010-09-19T23:56:16 Top 10 players: McLeopold(3601), kcm1700(3492), EBraun(3434), paulus(3399), adumlah(3355), golfin(3354), sighbot(3271), Titannet1(3245), locutus(3243), wagstaff(3223) 2010-09-19T23:56:36 Is anyone getting a "connection refused" message on the tcp server? 2010-09-19T23:56:51 McLeopold: btw, McLeopold.5 on tcp the bot you have on the main site? 2010-09-19T23:57:05 your main site one seems stronger 2010-09-19T23:57:30 No, it's designed to lose, but just barely. 2010-09-19T23:57:41 *** danielvf_ has quit IRC (Read error: Connection reset by peer) 2010-09-19T23:57:43 hah 2010-09-19T23:57:52 I learn more when I lose than when I win. 2010-09-19T23:57:53 McLeopold: on the official server each submission is considered a new bot and old submissions are simply hidden from the rankings 2010-09-19T23:57:57 ;) 2010-09-19T23:58:08 oh, and of course old submissions no longer play games 2010-09-19T23:58:16 Okay. I though I saw some odd ranking moves, so I was wondering. 2010-09-19T23:58:23 i also sit here just waiting for games I lose 2010-09-19T23:58:31 to analyse 2010-09-19T23:58:33 But, I can still see my games against old submitions. 2010-09-19T23:58:38 right 2010-09-19T23:58:48 because they still affect your current submission 2010-09-19T23:58:49 *** danielvf has joined #aichallenge 2010-09-19T23:58:51 McLeopold: if you want to be malicious, you need a bot which switches to random mode after x days :P 2010-09-19T23:59:08 lol 2010-09-19T23:59:13 that would be awesome 2010-09-19T23:59:25 what would that do for me? 2010-09-19T23:59:43 nothing, but it would screw up all the high ranked players that your bot already played against 2010-09-19T23:59:53 and beat, that is